Right policies can sustain productive 'creative destruction'

Innovation is built on a paradox. We need to reward those who create new technologies, but once they succeed, they may have an incentive to prevent the next generation of innovators from challenging them.

This tension lies at the heart of what economist Joseph Schumpeter called "creative destruction" — the process through which new innovations displace old technologies and render them obsolete. In the 1980s, Peter Howitt and I developed what is now known as the Schumpeterian growth paradigm to understand how this process drives economic growth.

The basic idea is that long-run growth comes from a cumulative process of innovation. Each innovator builds on what came before. Entrepreneurs invest in research and development because successful innovation can bring temporary economic rents. But those same rents can later be used to protect existing positions and discourage subsequent innovation.

This is why competition matters so much. Enterprises close to the technological frontier tend to respond to greater competition by innovating more, because they need to work harder to remain ahead. But those far from the frontier may react differently, becoming discouraged by stronger competition. A dynamic economy therefore needs to preserve both incentives for innovation and opportunities for new challengers.

This helps explain an important feature of the information technology revolution in the United States.

The IT revolution helped produce a major acceleration in US productivity growth from the mid-1990s to the early 2000s. At the same time, market concentration increased and "superstars" emerged. Companies such as Google, Microsoft, Amazon and Walmart were particularly successful at harnessing information technology and expanding their businesses.

Their rise initially helped drive economic growth. But as these companies grew increasingly dominant, they also discouraged new entrants to the market. The same technological forces that had helped accelerate growth eventually contributed to weaker competition and slower growth. In this sense, the history of the IT revolution illustrates the central contradiction of creative destruction: yesterday's innovators can become obstacles to tomorrow's innovation.

A similar challenge can arise when economies move from catching up to frontier innovation.

The Republic of Korea and Japan, for example, achieved rapid growth during periods when they were largely adopting and imitating frontier technologies. Europe similarly grew quickly after World War II by catching up with the United States. But imitation cannot drive growth indefinitely. Once an economy approaches the frontier, it needs an environment that encourages innovation at the frontier.

Large established corporate groups can make that transition harder by discouraging new entry and resisting stronger competition. That was part of the challenge posed by the ROK's chaebol and Japan's keiretsu, while Europe struggled more broadly to build an ecosystem conducive to frontier innovation.

There is an encouraging development today: China is performing strongly in frontier innovation. Its growing role in high-technology patenting shows that it is increasingly moving beyond simply adopting technologies developed elsewhere and becoming a source of innovation itself.

That makes the debate over artificial intelligence especially important.

Should we fear the AI revolution? I believe we should not, provided that we have the right institutions.

AI has enormous potential to raise productivity because it can automate tasks not only in the production of goods and services, but also in the production of ideas.

My colleagues and I estimate that AI-driven automation of production tasks could increase productivity growth by about 0.68 percentage points a year over a 10-year period. Once the effect of AI on the creation of new ideas is included, the potential gain could be even larger.

But AI also revives the competition problem.

The upstream parts of the AI value chain are already highly concentrated. Major technology companies dominate cloud computing, while the market for graphics processors is also highly concentrated. The lesson from the IT revolution is that a technology can have enormous growth potential while still failing to deliver sustained growth if competition is allowed to weaken.

Employment presents a similar paradox.

AI will replace some tasks, and some occupations could lose a substantial share of their existing work. But the overall effect on employment need not be negative. Companies that adopt AI can become more productive and competitive, increasing demand for their products and, in turn, hiring more workers. AI can also accelerate the discovery of new ideas, and new ideas generate new economic activity and new jobs.

The question, then, is not simply what AI will do. It is what societies will do with AI. To capture its growth potential, we need effective competition policy that allows new entities and new ideas to challenge established players. To capture its employment potential, we need strong education systems and labor-market institutions that help people move into new forms of work through retraining and income support.

The central lesson of creative destruction is therefore not that disruption is inherently good or bad. It is that economies must allow successful innovation to be rewarded without allowing yesterday's winners to close the door on tomorrow's innovators.

With the right institutions, creative destruction can generate not only faster growth, but sustained and inclusive growth. That is the promise — and the policy challenge — of the AI revolution.

The author is the winner of the 2025 Nobel Prize in Economic Sciences.

The views do not necessarily reflect those of China Daily.

If you have a specific expertise, or would like to share your thought about our stories, then send us your writings at opinion@chinadaily.com.cn, and comment@chinadaily.com.cn.