ThinkableWhat is this?

Globalisation & Trade

The Government That Bet on Steel — and Won (and the One That Bet on Batteries and Lost)

Every generation of policymakers believes it can spot tomorrow's winning industries — and every generation has both a POSCO and a Northvolt to explain.

The Idea

Industrial policy — the deliberate use of state funding, subsidies, and trade protection to favour specific industries — is one of the most contested ideas in economics. The sceptic's case is elegant: governments lack the information that markets aggregate across millions of buyers and sellers, so when they pick winners, they usually end up subsidising losers, locking in inefficiency, and inviting the industries they support to spend more energy lobbying than innovating. Research into China's Revitalization Plan for Ten Industries, a classic winner-picking strategy using subsidies and tax incentives directed at targeted sectors, found that it did not significantly enhance innovation quality — and may have actively discouraged it. The defender's case is equally powerful: some markets fail structurally. Advanced manufacturing has steep learning curves; a country that never builds its first factory never acquires the knowledge to build its second. Semiconductors, batteries, and shipbuilding are not like restaurants — you cannot simply let comparative advantage unfold when your rivals are themselves being state-subsidised. The IMF has tracked more than 52,000 industrial policy interventions across 75 countries since 2009, and the number of measures introduced in a recent year was roughly two and a half times the pre-pandemic average. Governments everywhere have already voted with their budgets. The honest answer is that industrial policy is not uniformly good or bad — it depends on what it is trying to fix, how it is designed, and whether the government in question can distinguish between supporting a genuinely stuck market and protecting a politically connected one.

In the World

Two cases, sitting on opposite ends of the ledger, make the tension vivid. In 1968, the South Korean government founded POSCO — the Pohang Iron and Steel Company — from scratch. The World Bank and the United States both explicitly declined to finance the project, judging that a state-run steel mill in a developing country like South Korea was economically unfeasible. The government went ahead anyway, funding construction partly from Japanese compensation payments for colonial rule, and placed the company under the leadership of retired Major General Park Tae-joon, who had already demonstrated an ability to turn around a struggling state-owned manufacturer. Within two decades POSCO had become one of the most efficient steel producers in the world, anchoring the country's transformation from an agrarian economy into an industrial one. The government partially privatised it via a stock listing in 1988, and full privatisation was completed in 2000. The bet paid off in ways the World Bank's models had not predicted, because the models could not easily value what economists call dynamic comparative advantage — the idea that you can build capabilities you do not yet possess. Now contrast that with Northvolt. Founded in 2015 as SGF Energy by two former Tesla executives, and rebranded as Northvolt in 2017, the Swedish battery maker was Europe's great green-industrial hope: a homegrown challenger to Chinese battery dominance, backed by billions in private capital and loan guarantees from the German government. By late 2024 it had delivered only a fraction of its planned production capacity. It filed for Chapter 11 bankruptcy protection in the United States in November 2024, then declared bankruptcy in Sweden in March 2025 — the largest bankruptcy in modern Swedish history, surpassing even the collapse of Saab Automobile. The technology and remaining assets of the Skellefteå factory were subsequently acquired by US battery startup Lyten.

Why It Matters

The lesson is not that governments should always back industries or never should. It is that the conditions for success are specific and demanding. POSCO worked partly because the South Korean state was tightly focused on export performance and had the discipline to exit — eventually privatising the company rather than protecting it forever. Northvolt failed partly because scaling a genuinely novel manufacturing process is brutally hard, and state backing could not substitute for operational execution. What should a curious, financially literate person take from this? First, be sceptical of both the reflexive free-market position (markets are always right, industrial policy is always waste) and the reflexive interventionist one (governments can steer the economy to better outcomes if they just try hard enough). The evidence is messier than either camp admits. Second, notice that industrial policy is not an abstract debate happening somewhere else. The CHIPS Act in the United States has directed tens of billions toward domestic semiconductor manufacturing, and as of 2026 that programme is still unfolding, with manufacturing milestones being hit even as the R&D side has stalled. These are choices being made right now, with public money, that will shape which industries exist in which countries for decades.

A Question to Ponder

If a government cannot know in advance which industries will succeed, is there a way to design industrial policy that makes the state a better learner rather than a better guesser — and what would that actually look like?

One reader's Thinkable for Finance Friday — a random draw from the library.

Start learning with Thinkable
Get your own draw, every day.Start free