The Return of Industrial Policy, and What It Quietly Walked Back In
Governments rediscovering the power of the state to direct investment have also, less visibly, rediscovered its power to pick winners.
The intellectual rehabilitation of industrial policy has been one of the more striking shifts in mainstream economic thinking over the past decade. Where state intervention in strategic sectors was once dismissed as inefficient at best and corrupt at worst, subsidy packages for semiconductors, clean energy, and critical minerals are now announced with bipartisan enthusiasm in capitals from Washington to Brussels to Seoul.
The shift is real, and in some respects overdue. But the speed of the rehabilitation has left important questions underexamined, questions that an older generation of development economists, burned by the failures of 1970s planning, might have pushed harder.
Who Decides What Is Strategic?
The category of 'strategic industry' has expanded with remarkable flexibility. Semiconductors clearly qualify. So, it turns out, does almost any manufacturing sector that employs enough people in enough marginal constituencies. The policy rationale and the electoral rationale are not always distinguishable from the outside, or, one suspects, from the inside.
The subsidy mechanisms themselves are frequently opaque. Direct grants, tax credits, accelerated depreciation, loan guarantees, regulatory carve-outs: the full value of a government support package is rarely disclosed in a single line. Journalists and researchers who try to calculate the actual public cost of a flagship investment deal often find the numbers scattered across multiple agencies and fiscal years.
The Leverage That Rarely Gets Mentioned
When a government offers substantial subsidy to attract a specific company to a specific location, it is, implicitly, accepting a significant shift in bargaining power. The company has been publicly courted. Walking away from the deal after announcement carries political costs that the initial commitment did not.
This is the leverage that rarely features in press releases celebrating inward investment. It tends to surface later: in renegotiated wage commitments, in environmental permit modifications, in the gradual softening of local content requirements. Industrial policy without robust conditionality and public audit is not a new idea so much as a very old one, the one that gave the concept a bad name the first time around.
More stories
How Long Until ChatGPT Cites You? Realistic GEO Timelines
Getting an AI assistant to name your brand isn't like ranking on Google, the clock runs on indexing, grounding, and model refresh cycles that rarely move at the same speed.
GEO Case Studies: What Measured Results Actually Look Like
Before judging any AI-visibility platform, it helps to know what a real result actually looks like, baseline, first citation, position, rather than a marketing chart.
Inside the New Business of Getting Brands Mentioned by AI Chatbots
As ChatGPT and Claude become a default place people ask "what should I buy," a service industry has formed around getting brands cited in the answer, here's how the legitimate end of it actually works.