Raymond Wang | Journal of Strategic Studies
ABSTRACT
Technological spillover – the transfer and use of knowledge between defence and civilian sectors – is desired by many countries, and particularly salient under intensifying US-China competition. Yet to date, no systematic measure of spillover exists. This paper presents a novel measure of spillover using a dataset of over 116,000 defence-related patents, and argues that (1) concerns over China’s ‘authoritarian advantage’ should be moderated, and (2) the composition and structure of knowledge transfer networks in the US and China should inform the appropriate channels of managing competition, which I demonstrate through case studies of aerial drones and missile technology spillover networks.
Introduction
Technological spillover – the transfer and use of knowledge between defence and civilian sectors – is a longstanding goal for governments. For instance, the US’s Defence Advanced Research Agency (DARPA) was created in part to encourage such knowledge transfers, resulting in canonical cases of spillover like the Internet.1 Both the US and China have pursued policies to encourage spillover. The US Department of Defence is trying to forge deeper ties with Silicon Valley, and created programs to help private companies navigate the Pentagon’s arduous acquisition process.2 The Chinese government has sought to encourage synergy between civilian and military sectors as early as the Deng era, and has elevated the policy into ‘military-civil fusion’ under Xi Jinping.3 Yet to date, there has not been a systematic attempt at defining and measuring spillover.
Being able to measure spillover is important for both scholarly and policy reasons. First, spillover is a way for countries to mitigate the guns versus butter tradeoff, the core issue in the political economy of security (PES).4 If advancements in defence can increase productivity in the civilian economy through spillover, then a dollar invested in guns will no longer be a dollar less for butter; instead, it could catalyse innovation in both guns and butter, thus mitigating the tradeoff. How states try to circumvent this tradeoff is central to the PES subfield.5
For policymakers, technological competition is a central component of US-China competition.6 Increasing productivity through spillover is a policy goal for both countries, with recent research assessing each country’s industrial strengths and vulnerabilities.7 Yet to date, a systematic measure of spillover is lacking.
Furthermore, politicians speculate that China may have an ‘authoritarian advantage’ by ‘mobilizing its bureaucracy and assigning ambitious targets’ with a ‘single-minded determination that democracies lack.’8 Given that spillover is one important facet of technological competition, measuring it can help assess whether such claims are justified. To preview the paper’s findings: while China can generate knowledge flows, inefficiency problems mean that claims about China’s ‘authoritarian advantage’ need qualification. Moreover, if current trends persist, the US is poised to maintain a lasting advantage, especially in its defence sector’s capability to draw upon its advanced civilian technical base.
Finally, mapping out spillover networks allows scholars to explore the characteristics of different technological sectors across countries. Two attributes in particular are relevant to managing technological competition: who is involved in a knowledge network and its structure. Identifying the key players in a network informs the appropriate channels for arms control discussions, while network concentration affects how arms control proposals are received domestically.
This article proposes a replicable, open-source methodology that measures spillover across countries using patent citation data. All patent data in this paper is drawn from PatSnap.9 In brief, I construct a novel dataset of 116,097 defence-related patents, representing all active defence patents published between 2012–2022. I then identify the civilian patents that these defence patents cite (backward spillover), and those that cite the defence patent in the future (forward spillover). Finally, I filter for spillover that occurs within the same country and between entities of that country, what I call internal spillover. I show that the measure is noisy but captures processes that fit an intuitive understanding of spillover. As such, it offers a proof-of-concept to catalyse a research agenda on spillover.
The rest of the paper proceeds as follows. The first section defines spillover in more detail and justifies the focus on internal spillover. The second section explains the measurement strategy and its potential shortcomings. The third section presents summary statistics of cross-national performance and shows preliminary evidence that echoes the findings of scholarship on China’s innovation system: it can generate scale but faces efficiency challenges.10 The fourth section conducts case studies of spillover networks in aerial drones and missile technology in the US and China. The final section invites scholars to contribute to this research agenda by identifying areas for further research.
[Read full article here] or [PDF Version]
From Journal of Strategic Studies