Jacob Schindler

The question of whether China’s IP system privileges domestic companies has provoked plenty of debate on this blog and elsewhere in the past. Most recently, we reported last week that non-Chinese plaintiffs in civil lawsuits at the Beijing IP Court had a perfect 65-0 win rate last year, a finding that surprised some of the readers we heard back from. A more rigorous study published in Issue 76 of IAM found an overall 70% win rate for foreign companies in Chinese IP suits, concluding that the system “often benefits foreign interests at the expense of domestic ones”. Enforcement, though, is only possible if you have a granted patent in the first place. And a new study suggests that the patent examination process may present what the authors call a “subtler barrier” to overseas corporates.

The authors of the study (which can be downloaded in full here), Gaétan de Rassenfosse and Emilio Raiteri of the Ecole Polytechnique Fédérale de Lausanne, set out to find out whether the probability of being granted a SIPO patent differed for domestic and foreign applicants, all other factors being equal. They were particularly interested in technologies identified in China’s industrial policy as strategically important. These areas include, among others, information technology, biotechnology, energy, advanced materials and advanced manufacturing. The sample comprised nearly half a million patents and controlled for the quality of the applications in a variety of ways laid out in the paper.

The headline finding is this: “Results suggest no – or at worst only weak – overall discrimination against foreigners at SIPO. However, foreign patent applications in ‘strategic’ technology areas are about four to seven percentage points less likely to receive a patent grant than similar domestic applications.” Given the importance of industrial policy and central planning in the Chinese system, the authors say their results suggest “a case of technology protectionism by means of the patent system”.

The researchers found a statistically significant bias in three of the eight key sensitive fields: biotechnology, ICT and advanced energy. The effect was the strongest in biotech – the probability of a SIPO grant for a non-Chinese entity was “between 14 and 15 percentage points lower” than what the model would have expected. In ICT and energy, the negative effect was between 5% and 10%.

Included in the sample were about 2,132 patents declared as standards essential – 88.5% of which related to ETSI-developed standards. Summing up their findings the authors write: “Applications by foreigners related to SEPs have a particularly low probability of being granted, with an anti-foreign bias ranging from 9 to 14 percentage points.”

There are various explanations for why success rates may be different for domestic and foreign patent applicants in any country. Better knowledge of the local IP system is one, especially where there is a language barrier. But de Rassenfosse and Raiteri did control for the quality of the patent attorney firms responsible for translation and filing of the patents. Another is that firms may be more likely to push for a grant if they come up against a prosecution hurdle in their home jurisdiction. But if either of these factors were at play in a major way, you’d expect them to affect the entire sample, not just the more sensitive technologies. The authors point out that there is no evidence of organised or intentional discrimination at SIPO, but insist that the reasons for the discrepancies in grant rate “are not to be found among the applicants or the attorneys”.

Intriguingly, some observers have suggested that SIPO assigns more experienced examiners to applications by foreigners in strategic areas, which may account for the gap in grants. That would be ironic, given that this policy could be intended to ensure high standards where foreign parties are involved. But de Rassenfosse stresses that even if this were the case, it would still violate the principle that foreign applicants should receive equal treatment to domestic ones.

It certainly seems like a line of inquiry that is worthy of further investigation, and I suspect SIPO will be as interested as any in seeking an explanation for these results (though, maybe not in public). One positive is that SIPO application and grant data is easier to collect in large samples than data on Chinese litigation, so there should be plenty of opportunity for further analysis. De Rassenfosse says he plans to look more specifically at SEPs in future work.

In the meantime, I don’t expect this notion of potential bias to curb the enthusiasm of foreigners for Chinese patents; if anything, it may warrant filing more applications per invention to ensure protection.