Korean Industrial Policy

In The You On AI Field Guide

The Korean transformation defies the conventional development framework on multiple dimensions. The country lacked obvious comparative advantages — limited natural resources, devastated infrastructure, no significant industrial base, no strong educational system. By orthodox theory, Korea should have specialized in low-value-added activities reflecting its existing comparative advantage. Instead, it deliberately built industries in which it had no current comparative advantage but for which strategic state action could produce future competitive capability.

The success was not inevitable. Park Chung-hee's regime was politically authoritarian, several of its industrial bets failed (the heavy and chemical industries push of the 1970s required substantial restructuring), and the model produced significant inequality and labor repression alongside its economic gains. But the overall trajectory — from fish and tungsten exports to semiconductors and automobiles — is unmistakable, and the mechanism is unmistakable: strategic state intervention on a scale that contemporary orthodoxy would denounce as massively distortionary.

The relevance to AI is structural. The Korean experience demonstrates that nations can move from technological periphery to technological frontier through deliberate, sustained, well-designed industrial policy. The contemporary AI industry's geographic concentration in a handful of nations is not a permanent feature of the technology — it is a temporary configuration that strategic action by other nations could substantially reshape. The barrier is not technical or even economic. It is political, institutional, and ideological — the constraints imposed by the contemporary international order on the policy tools that successful developers historically used.

Chang's analytical contribution has been to extract from the Korean case the institutional design principles that made the policy effective: competent bureaucracy insulated from short-term political pressure, performance-conditioned subsidies that disciplined private firms, integration of policy across sectors and time horizons, and the strategic patience to invest in capabilities whose returns would not appear for decades. These principles can be applied by any nation willing to make the institutional investment — the obstacle is not capability but political will and international policy space.

Origin

The Korean industrial policy framework was implemented under General Park Chung-hee, who took power through military coup in 1961 and was assassinated in 1979. The framework was elaborated by the Economic Planning Board, established in 1961 as the principal coordinating agency, and implemented through a network of state-controlled banks, sectoral ministries, and the chaebol industrial groupings that became Korea's national champions.

The intellectual influences on Korean industrial policy included the Japanese MITI model, the German developmental tradition, and direct study of Hamilton's American industrial policy program. The implementation drew heavily on Korean technocrats trained at American universities and exposed to American developmental economics — a generation of policymakers who learned from American academic theory the lessons that contemporary American policy advice now denies.

Key Ideas

Strategic capability building. Building industries in which the country has no current comparative advantage but for which strategic state action can produce future competitive capability.

Coordinated toolkit. Directed credit, import protection, technology transfer requirements, export targets, education investment — deployed in coordination across multiple decades.

Performance discipline. Subsidies and protection conditional on meeting export targets, quality benchmarks, and technology acquisition milestones — preventing the slide into rent-seeking.

Empirical refutation. The most successful single case of developmental intervention provides paradigmatic evidence against the orthodox claim that strategic state coordination does not work.