Yang Yanqing | Technological decoupling or a "small courtyard with high walls"? — An empirical analy

Yang Yanqing,Director of the Planning and Development Department and Executive Head of the Digital Economy Research Team at the Shanghai Artificial Intelligence Laboratory.He is a Senior Research Fellow at the National Institution for Finance and Development, a Visiting Professor at the School of Economics, Fudan University, a Standing Committee Member of the 4th Youth Committee of the All-China Federation of Returned Overseas Chinese, a Member of the 13th Shanghai Municipal Committee of the Chinese People's Political Consultative Conference, a Member of the Shanghai Private Economy Development Strategy Advisory Committee, a Member of the Sino-Singapore (Chongqing) Financial Expert Advisory Group, and a Director of the Fudan University Alumni Association. He holds Bachelor's, Master's, and Doctoral degrees in Economics from Fudan University and was a Visiting Scholar at Johns Hopkins University. He is one of the core creative teams of *First Financial Daily*, China's first market-oriented financial daily newspaper, and has created several high-end financial dialogue television programs, engaging in dialogues with over a hundred central bankers, finance ministers, political figures, business leaders, economists, and scientists worldwide. He has served as a speaker and host at top-tier domestic and international forums such as the World Economic Forum, the IMF Annual Meetings, the Boao Forum for Asia, the China Development Forum, the Lujiazui Forum, and the World Artificial Intelligence Conference. He led the transformation of the First Financial Research Institute into an economic policy think tank, which was selected as one of the first batch of key think tanks in Shanghai. He has long been involved in tracking and actively researching macro-finance, the digital economy, and global governance, publishing numerous papers in authoritative academic journals, presiding over several provincial and ministerial-level research projects, and publishing dozens of books in Chinese and English. Recent research focuses on artificial intelligence and innovation economics.

Yang Yanqing

Based on global invention patent big data, this paper uses machine learning to empirically study the heterogeneous impact of the technological friction in the first phase of the Sino-US trade conflict on different technological fields in China, and explores the characteristics of the technological fields affected by different impacts. Since the impact of Sino-US technological friction on different technological fields in China varies, this paper uses the C-Lasso machine learning method and the DID causal inference framework. Taking the Section 301 investigation announced by the US government in March 2018 as the policy time point, it systematically characterizes the impact of technological friction on different technological fields based on two classification criteria: strategic emerging industries grouping and over 30,000 micro-level IPC groupings. The grouping results of the C-Lasso method show that the technological fields affected in China are limited to certain specific areas: 13 out of 40 strategic emerging industries were significantly affected, and the affected groups accounted for 3.3% of the 31,851 IPC coding groups. This is consistent with the characteristics of "small courtyards and high walls" rather than "technological decoupling".

Su et al. (2016) proposed a penalty function-based machine learning method to identify latent group structures in panel data and found that panel regression in heterogeneous data often exhibits strong parameter sparsity, a characteristic that fits well with the Lasso method in machine learning. This paper uses the C-Lasso method to identify latent group structures in panel data, employing three sets of patent data (US patent application data from the China Patent Office; US patent application data from its home patent office (USPTO) and invention patent citation data from seven major patent-granting countries worldwide; and US patent citation data for basic scientific papers), fully considering the lag factor of patent applications.

This article uses the C-Lasso machine learning method and the DID causal inference framework, taking the Section 301 investigation announced by the US government in March 2018 as the policy time point. Based on two classification criteria—strategic emerging industries grouping and over 30,000 micro-level IPC groups—it systematically characterizes the impact of technology friction on different technology sectors. The C-Lasso method's grouping results show that the impact on China's technology sectors is limited to certain specific areas: 13 out of 40 strategic emerging industries were significantly affected, and 3.3% of the 31,851 IPC coding groups were affected, consistent with the characteristics of "small courtyards and high walls" rather than "technology decoupling." The empirical conclusions of the DID framework show that in the first phase of the Sino-US trade conflict, the number of US invention patent applications in China within the "small courtyards and high walls" category persistently decreased by 15.6%.

The technological fields impacted by the US-China technology trade show exhibit three characteristics: first, the US has a stronger comparative advantage in technology; second, the gap between China and the US is rapidly narrowing; and third, there is a greater reliance on scientific advancements in these areas. The Sino-US technology friction has disrupted and damaged normal technological exchanges between the two countries, and its impact on global and Chinese technological innovation will be significant and far-reaching. China's accelerated process of achieving technological self-reliance and strength cannot be separated from the development of basic scientific research. Addressing the shortcomings in basic scientific research as soon as possible and forming a new, virtuous cycle of scientific-driven technological innovation is also an urgent priority.

Li Xin

School of Statistics, Beijing Normal UniversityProfessor and Vice DeanLi XinThis article is considered to be highly research-oriented and policy-driven. The study provides a more detailed structural analysis of the impact of US-China technological friction, and uses the number of US patent applications in China to study the heterogeneous impact of the US Section 301 investigation on different technological fields in China, thus revealing the true situation of Sino-US technological friction that cannot be reflected by total data.

This article can further explore this issue from the following two aspects: First, the article could consider incorporating a cross-departmental context. The research approach is to anchor the areas of Sino-US technological friction by examining changes in US patent applications in China and directly measure the impact of this friction. Due to the lag in patent applications, there is a lack of discussion on the impact of Sino-US technological friction since 2020. Could we use IPC information to match micro-products, such as customs data, to examine the impact of Sino-US technological friction on the trade of intermediate goods between the two countries, which reflects technological connections? Second, could the article incorporate the alliance effect? ​​Unlike the Trump era since 2020, the US now adopts a strategy of competition + investment + alliances. If it were just the US alone, the total impact of technological decoupling from China might not be significant. However, if the alliance effect is considered, the total impact could be substantial. Especially given the US's continuous use of various controls to build "walled courtyards," promoting "decoupling and supply chain disruptions" while simultaneously pushing for "friendly outsourcing" and establishing a "chip alliance," considering the alliance effect is even more important.