CLM Insights Interview with Jeffrey Ding
Jeffrey Ding. Technology and the Rise of Great Powers: How Diffusion Shapes Economic Competition. Princeton University Press, 2024. 320 pp. ISBN-10: 0691260346; ISBN-13: 978-0691260341
Insights Interview
Can you explain briefly why diffusion of a breakthrough general-purpose technology (GPT) is so important to shift the balance of power among nations?
Technologies only make their mark when they diffuse to a population of potential users. GPTs, like electricity, are special because they represent foundational advances that can be used in a variety of ways, which means the population of their potential adopters encompasses nearly all economic sectors. Thus, a nation that effectively adopts a GPT throughout its entire economy can achieve a significant boost in productivity, which is a key determinant of long-term economic growth.
Understanding that not all technologies are created equal – instead of treating all emerging technologies with the same boilerplate language – points the way toward diffusion as the most important phase of technological competition when it comes to GPTs. It is difficult for one nation to corner all the innovations in foundational technologies, such as electricity or computers. The impact of GPTs materializes via the dissemination of the GPTs across different productive processes in the economy. The central argument in my book is that when one looks at how past technological revolutions affected the rise and fall of great powers, the country that became the next leading economic power was the one that succeeded in diffusing GPTs across its entire economy.
What determines or influences a country’s success in diffusing a breakthrough GPT?
A country’s ability to capitalize on technological revolutions depends on its adaptations to the demands presented by GPTs. Focusing on skill formation institutions, I emphasize the importance of education and training systems that broaden the base of engineering skills connected to a GPT. These institutions not only address talent shortfalls but also help the GPT and application sectors coordinate development by sharing information – both effects enable the spread of a GPT.
This pathway differs significantly from the “leading sector” theory, which fixates on a country’s capacity to monopolize innovations in particular fast-growing industries. When it comes to skill formation institutions, the “leading sector” theory highlights the significance of cultivating heroic inventors and boosting cutting-edge R&D.
Why do countries fail to translate their advantages in a “leading sector” (LS) into enduring economic power? Based on your theory, it seems that China’s current advantages in clean energy and EVs will unlikely be a game-changer in its economic trajectory. Is that correct?
Correct. I think the current hype about China’s advantages in electric vehicles falls into the “leading sector” trap. Electric vehicles may have a significant impact on the transportation sector and have ramifications that extend beyond economic benefits (e.g., decarbonization and environmental impacts). However, when it comes to game-changers for long-term economic trajectories, electric vehicles are not GPTs.
One way to illustrate why AI is a more general-purpose technology than electric vehicles is to consider the simple fact that advances in AI are transforming the electric vehicle sector (e.g., by enabling predictive maintenance and providing more reliable estimates of energy demand for smart grid integration), whereas advances in electric vehicles are not transforming the field of AI. Because GPTs have a wide variety and range of uses, they hold the promise of significant productivity gains. As Nobel Prize–winning economist Paul Krugman once said, “Productivity isn’t everything, but in the long run it is almost everything.”
In your book, you cite Japan as a cautionary tale. What does the Japanese experience tell us about the obstacles China will face in the “4th Industrial Revolution”?
In the late 20th century, fundamental breakthroughs in information and communication technologies presented another opening for a shift in economic leadership. Many commentators, including some of the most prominent thinkers and policymakers at the time, warned that Japan’s lead in semiconductors and consumer electronics, supported by its keiretsu system of industrial organization and its aggressive industrial policy, would threaten U.S. economic leadership. However, despite its advantage in leading sectors, Japan’s productivity growth stalled in the 1990s. In part, this was because Japan did not lead the U.S. in the diffusion of general-purpose information technologies.
One of the most important findings from the chapter on U.S.-Japan competition is that in sectors that produced information and communications technologies (ICT), Japan’s total factor productivity growth kept pace with that of the United States; however, in sectors that intensively used IT, Japan’s total factor productivity growth lagged far behind. The lesson here is that it is really important to accurately assess China’s technological capabilities so to avoid generalizations from a few exceptional sectors that may lead analysts to overlook developments in the struggling sectors.
In the unfolding U.S.-China AI competition, what are the current and potential blind spots for both sides?
For China, the more that it falls back into a party-led development model, the less effective it will be in adopting AI at scale. While top-down, centralized approaches may be helpful to encourage the widespread adoption of certain types of technologies such as high-speed rail, these tactics are much less effective for GPTs since their diffusion to small and medium-sized businesses must be driven by organic, market-based forces. The historic example of the Soviet Union underscores this point. While it was able to pioneer new innovations, it lacked the fast-acting diffusion process of decentralized, market-based institutions.
As for the U.S., it is overly focused on preventing cutting-edge innovations from leaking to rivals including China (the leading-sector model). For example, the logic behind U.S. export controls on advanced AI chips is a vision of technological leadership that is rooted in monopolizing innovations. This vision goes against the historical evidence I have gathered in this book from the past three industrial revolutions. Instead, the U.S. should prioritize policies and investments that spur the widespread adoption of AI across the entire economy.
