Bias in the Bureaucracy: The US-China Technology Conflict and Patent Examination

Abstract

We study whether US patent examiners, who apply rules that make no distinction by origin, treated Chinese applications differently during the US-China technology conflict. No policy measure in the conflict addressed patent examination, but examiners choose how deeply to search the prior art and how strictly to judge a marginal claim. We estimate a triple difference model with examiner and technology-by-filing-cohort fixed effects on 1.34 million applications that list an inventor in China or in one of three comparator economies (South-Korea, Taiwan, or Japan). After the Department of Justice launched its China Initiative in November 2018, the China-comparator gap in rejection rates at the first office action widened 1.73 percentage points more in sensitive technologies (those that Made in China 2025 prioritized) than in other technologies. Relative to comparators, the rejection rate for Chinese applications rose by 0.86 points in sensitive technologies and fell by 0.88 points elsewhere. Changes in the composition of applicants largely explain the fall but not the rise. We detect no widening before the Initiative, including during the 2017-2018 trade escalation. Measured by year of first action, the widening was close to zero in 2019, about one point in 2020-2021, and two to three points from 2022 onward. The widening appears in prior-art rejections and is absent from rejections for ineligible subject matter. We interpret the widening that composition does not explain as bureaucratic bias.