How Governments Can Succeed With Industrial Policy Through Directed Experimentation

Here is your publish-ready WordPress Gutenberg block HTML article based on the verified Project Syndicate source, adhering strictly to the rules provided:

Governments worldwide are increasingly embracing industrial policy as a tool to navigate economic uncertainty—but the challenge lies not in whether to intervene, but in how to do so effectively when outcomes remain unpredictable. The answer, according to a new analysis, lies in directed improvisation: a framework that fosters experimentation in targeted sectors rather than relying on rigid, top-down planning.

The concept gains urgency as nations grapple with shifting global supply chains, technological disruption, and the lingering effects of the COVID-19 pandemic. China’s Made in China 2025 initiative, for instance, exemplifies both the promise and pitfalls of state-led industrial strategy. While the program has accelerated domestic innovation in advanced manufacturing and artificial intelligence, its heavy-handed approach has also sparked trade tensions and raised questions about overreach.

Why Traditional Industrial Policy Fails in Uncertain Times

The core dilemma, as outlined in the analysis, is that governments often assume they can predict which industries will thrive in the future. Yet in an era of rapid technological change—where breakthroughs in AI, biotech, and quantum computing reshape economies overnight—such forecasts are inherently unreliable. The solution? Creating environments where experimentation is encouraged, failures are tolerated, and successful models can be scaled.

This approach contrasts sharply with historical industrial policies that relied on centralized planning. For example, South Korea’s rapid ascent in the 1970s–1990s was fueled by targeted state support for chaebols like Samsung and Hyundai, but even those successes depended on adaptive governance. Today, the task is more complex: governments must balance support for legacy industries with bets on unproven technologies, such as next-generation semiconductors or carbon-capture methods.

Directed Improvisation: How It Works

The framework hinges on three pillars:

• Targeted experimentation: Governments should identify high-potential sectors—such as green energy, AI-driven healthcare, or e-commerce platforms—and fund pilot projects with clear exit criteria. For instance, Singapore’s National Research Foundation has successfully deployed this model by backing startups in precision medicine, where outcomes are inherently uncertain.
• Decentralized decision-making: Rather than dictating outcomes, policymakers should empower local governments, universities, and private sector consortia to test solutions. Germany’s dual education system, which blends vocational training with industry partnerships, is a case study in how decentralized innovation can outperform top-down mandates.
• Real-time feedback loops: Data-driven monitoring—leveraging AI and public-private partnerships—allows governments to pivot resources toward what’s working. Estonia’s digital governance model, where policy adjustments are made in near real-time based on citizen feedback, demonstrates this principle.

Critics argue that such flexibility risks inefficiency, particularly in sectors requiring massive upfront investment (e.g., semiconductor fabs or nuclear fusion). However, the analysis counters that the alternative—bet-the-farm commitments to losing propositions—is far costlier. For example, the U.S. Semiconductor industry’s resurgence under the CHIPS and Science Act reflects this balance: while the $52 billion in subsidies is substantial, We see distributed across multiple firms and use cases, reducing systemic risk.

China’s Made in China 2025: A Cautionary Tale and a Case Study

China’s industrial policy offers a mixed blueprint. On one hand, the initiative has propelled the country to the forefront of electric vehicles, solar panels, and 5G infrastructure. State-backed firms like BYD and Huawei now dominate global markets in these areas, proving that targeted intervention can yield results. Yet the program’s heavy reliance on forced technology transfer and protectionist measures has alienated trading partners and triggered retaliatory tariffs, undermining its long-term viability.

The lesson? Effective industrial policy must align economic goals with geopolitical realities. China’s success in renewables, for instance, stems from its ability to combine state direction with market mechanisms—such as allowing private firms to compete in solar panel manufacturing while maintaining state control over critical minerals supply chains. By contrast, its struggles in high-end semiconductors (despite years of subsidies) highlight the limits of command-economy approaches when facing entrenched global competitors like TSMC.

The Role of E-Commerce and AI in Shaping Policy

Two sectors illustrate the tension between state intervention and market dynamics: e-commerce and AI. In China, platforms like Alibaba and JD.com were initially nurtured through regulatory sandboxes that allowed rapid experimentation with logistics and fintech innovations. However, as these firms grew dominant, the state shifted to tighter controls—reflecting the challenge of balancing innovation with antitrust concerns.

Similarly, AI presents a paradox. While governments worldwide are racing to fund AI research (e.g., the EU’s €1 billion Digital Europe program), the technology’s decentralized nature makes it resistant to traditional industrial policy tools. The U.S. Approach—combining federal grants with private-sector-led consortia like the Partnership on AI—offers a model for how directed improvisation can work in this space. By focusing on ethical guidelines and infrastructure (e.g., cloud computing) rather than picking winners, policymakers reduce risk while fostering ecosystem growth.

What Comes Next: Lessons for Governments

The analysis concludes that the future of industrial policy lies in agile governance: systems that can adapt as quickly as the technologies they seek to shape. Key steps include:

• Pilot programs over monolithic projects: Small-scale experiments in emerging fields (e.g., lab-grown meat or space tourism) allow governments to learn without overcommitting.
• Public-private innovation hubs: Models like Israel’s Yozma program, which matched venture capital with state funding, show how collaboration can de-risk high-potential bets.
• Global collaboration where possible: The International Energy Agency’s work on hydrogen standards demonstrates that even in competitive sectors, shared frameworks can reduce duplication.
• Transparency in failure: Governments must be willing to publicly acknowledge and learn from missteps, as Japan did after its “lost decades” revealed the flaws in its post-bubble industrial strategies.

The stakes could not be higher. As geopolitical fragmentation deepens and technological disruption accelerates, the ability to experiment without fear of failure may well determine which economies thrive—and which fall behind. The question is no longer whether governments should intervene, but how wisely they can do so.

This article is based on analysis from Project Syndicate, published May 25, 2026.