Chaebols and firm dynamics in Korea

GAICIL Insights

Sometimes the best thing a crisis does is humble a giant.¹ Chaebols and firm dynamics in Korea, by Philippe Aghion, Sergei Guriev and Kangchul Jo, studies what happened when the 1997-98 Asian crisis and subsequent reforms reduced the dominance of Korea's chaebols. Large business groups can drive industrialization and can also entrench themselves at the expense of new competitors. The episode provides a rare natural experiment on loosening incumbent dominance.

The findings are striking. In industries dominated by chaebols before the crisis, labor productivity and total factor productivity of non-chaebol firms rose markedly after the reforms, relative to other industries. The TFP increase was especially large for non-chaebol firms in previously chaebol-dominated industries. Entry of non-chaebol firms increased significantly in all industries after the reform, while chaebol markups declined substantially, especially where chaebols had dominated. Non-chaebol firms also dramatically increased their patenting. The shade lifted, and the undergrowth shot up.

The authors read this as Korea moving from catching-up growth based on investment in existing technologies to innovation-based growth. By weakening chaebol power, the crisis opened a window for reforms that increased competition, and new and smaller firms seized it, raising productivity and innovation. The giants built the economy that the challengers then took to the frontier.

The parallel with artificial intelligence is hard to miss. AI markets feature large, diversified technology groups with significant positions across many products and services. The Korean experience suggests that when dominant incumbents' advantages shrink, independent firms flourish, entering markets, raising productivity and innovating more. Dynamism depends on space for challengers.

The development-stage lesson matters most. Economies at the technological frontier must rely on innovation, and innovation thrives on competition and entry. Large groups can mobilize investment effectively to catch up. Sustaining frontier growth requires a competitive environment where new firms can challenge established ones. For economies seeking leadership in AI, keeping markets open to entrants is a strategic necessity. What wins the catch-up race can lose the frontier race.

Falling markups widen the gains. Lower markups mean more competitive pricing, benefiting consumers and downstream firms. Reforms reducing entrenched market power can deliver on several fronts at once: lower prices, more entry, higher productivity and more innovation.

The crisis was the catalyst, and policymakers need not wait for one. It weakened incumbents and created political space for reform. Proactive measures ensuring contestability in key markets, including AI, can produce similar benefits without a crisis's costs. Reform by choice is cheaper than reform by collapse.

Industrial policy gets a timing rule. Supporting large national champions may help at some stages of development. The Korean evidence indicates that as economies approach the frontier, the benefits of competition and entry grow. AI policies supporting research and adoption while keeping markets open to challengers may best promote innovation. National champions have an expiry date, and it arrives near the frontier.

The identification is careful. Comparing industries with different degrees of prior chaebol dominance isolates the effect of reduced incumbent power on other firms, a reliable basis for policy lessons.

The patenting surge is the heart of the story. Independent firms, freed from constraints imposed by dominant groups, invested in new technologies. Entry and competition translated into innovation, exactly the engine of long-run growth.

The thread running through it all: reducing entrenched dominance can unleash productivity and innovation. When Korea's chaebols lost some of their advantages, independent firms entered, became more productive and innovated more, while chaebol markups fell. As AI reshapes the global economy, ensuring dominant firms do not foreclose opportunities for challengers will be essential for innovation-based growth.

Here is the uncomfortable test for governments courting AI national champions: the Korean record suggests the champions may be most valuable once their grip loosens. A country that wants AI startups to patent like Korea's post-crisis challengers may need to open space for them before any crisis forces the issue.