By Vishnu Venugopalan
Two landmark publications on industrial policy for development arrived within two weeks of each other: a World Bank policy research report on Industrial Policy for Development approaches for the 21st century, and a Dani Rodrik–Mariana Mazzucato paper on the taxonomy of conditionalities. Together, they offer the most substantive framework for designing and governing industrial policy that policymakers can no longer afford to ignore.
In 2019, the International Monetary Fund (IMF) published a paper with an unusually candid and provocative title: The Return of the Policy That Shall Not Be Named. It said plainly what most institutions had spent decades refusing to admit — that industrial policy was practiced everywhere, acknowledged almost nowhere, and treated by the Washington Consensus as the road to perdition
Since then, discussions around industrial policy in practice have picked up pace — in capitals, at development banks, in academic journals, and in corporate boardrooms. Two publications released within two weeks of each other bring that momentum into sharp focus. The first is the World Bank's new Policy Research Report, Industrial Policy for Development: Approaches in the 21st Century, by Ana Margarida Fernandes and Tristan Reed, the institution's most comprehensive reckoning with the subject in a generation. The second is a peer-reviewed article in Industrial and Corporate Change by Mariana Mazzucato and Dani Rodrik, "Industrial policy with conditionalities: a taxonomy and sample cases", a practical framework for turning public investment into genuine public value. Read together, a compelling picture emerges, theory is finally catching up with practice, and practice desperately needs the theory. The question is no longer whether governments should do industrial policy. It is whether they know how and whether their institutions are built to serve the public, not capture it.
World Bank’s Road to Perdition to Policy Toolkit
Thirty years ago, the World Bank was among industrial policy's sharpest critics. The 1993 East Asian Miracle report concluded that selective interventions "generally did not work" and held "little promise for other developing economies." That verdict cast a long shadow over development policymaking for a generation. The new report's foreword dispenses with any pretense of continuity: "That advice has not aged well — it has the practical value of a floppy disk today. The observations are stark. Among 183 countries surveyed, every single one targets at least one industry in its national development plan. Business subsidies in upper-middle-income economies now average 4.2% of GDP, the highest on record. In a recent poll of World Bank country economists, 80% reported that client governments were actively seeking advice on industrial policy. The question is no longer whether governments do industrial policy. It is whether they do it well.
The Bank's answer is a refreshingly practical framework. It maps 15 policy tools ranging from industrial parks and skills development to production subsidies, export bans, and competitive exchange rate interventions and matches their feasibility to three country characteristics: local market size, government bandwidth, and fiscal space. A small, low-capacity economy with limited fiscal room cannot realistically deploy technology-transfer quid pro quos or sophisticated consumer demand subsidies. What it can do is invest in industrial parks or quality infrastructure, which are blunter instruments, but achievable ones. The implicit message matters: design your policy ambition to fit your institutional reality, not your aspirations. The evidence the report marshals is persuasive. Romania became a global software hub through targeted payroll tax exemptions for qualified engineers. Brazil reshaped its agriculture through redirected public research. South Korea's heavy and chemical industry push caused GDP to be an estimated 3% larger annually in the long run — a benefit that dwarfs the upfront subsidy cost the Bank's own 1993 report had flagged as prohibitive. Industrial policy, the Bank now concedes, is not only replicable across contexts. In many of them, it is necessary.
The Missing Ingredient: Conditionality
Here is where the World Bank report, for all its ambition, leaves a gap — and where Mazzucato and Rodrik’s paper steps in with precision. The Bank's report is strong on what tools to deploy and when, given a country's characteristics. But it is thinner on how to structure the deal between governments and firms in a way that actually generates public value, not just private profit dressed up in policy language. Subsidies without conditions are handouts. With the right conditions, they become transformative deals. Conditionality is the mechanism that turns public spending into public value — embedding social, environmental, and economic obligations into the very structure of public-private contracts. That is exactly the terrain the Mazzucato–Rodrik taxonomy inhabits.
Their argument begins from a simple but underappreciated observation: some measure of conditionality is inherent in any industrial policy. Public support is always provided in return for something. The question is whether that "something" is made explicit, coherent, and oriented toward public purpose — or left vague, implicit, and easily captured. The taxonomy they develop organizes conditionalities along four analytical dimensions:
The taxonomy is not an abstract exercise. Mazzucato and Rodrik ground it in nine case studies spanning Germany, the UK, the US, Israel, Italy, and South Korea — covering green energy, semiconductors, pharma, regional development, and heavy industry. The lessons are rich, sometimes uncomfortable. Germany's KfW energy efficiency programs demonstrate what fixed, well-monitored conditionality can achieve: for every euro invested, the state earned roughly four euros back through taxes, reduced unemployment spending, and social contributions — while generating 64,000 full-time jobs and cutting 700,000 tons of CO₂ annually. The Oxford–AstraZeneca partnership shows how a procurement-linked non-profit condition, combined with advance purchase guarantees, produced 2.6 billion vaccine doses at accessible prices to 170 countries. Israel's R&D incentive system demonstrates royalty-based profit-sharing, which means firms only pay back when profitable, maintaining genuine risk-sharing without chilling innovation.
But the cases also carry warnings. South Korea's chaebol system, built on powerful sectoral conditionality that drove extraordinary growth, created a dependency between state and conglomerates that spawned monopolistic excess, political capture, and recurrent bailout cycles. Italy's Law 488/92 regional subsidies, well-intentioned in design, were compromised in execution by clientelism and, in some cases, organized crime. Scotland's ScotWind offshore wind leasing secured major investment commitments but set price caps that, by some estimates, left up to 18 times more potential revenue on the table than comparable UK and US programs captured.
The pattern is quite consistent. Conditionality works when it is clear, monitorable, and enforced. It fails when it is vague, when monitoring is absent, or when the government lacks the institutional capacity to hold firms to account.
The Institutional Imperative
This is where both publications converge on a shared blind spot in the broader policy conversation: state capacity is not optional. It is the load-bearing infrastructure beneath every tool in both frameworks. The World Bank is candid about this. Government "bandwidth", possibly an alternate term for State Capacity, in this context, the capacity to interact with many firms and industries simultaneously, to design credible criteria, and to monitor outcomes, is one of the three decisive variables determining which industrial policy tools are feasible. The Bank points to Peter Evans' concept of "embedded autonomy": the most effective industrial policy agencies combine deep ties to the private sector with genuine independence from capture by it.
Mazzucato and Rodrik are equally explicit. The design of conditions is a delicate task. Too many micromanaged requirements stifle innovation and entrepreneurial discovery. Too few, and subsidies become unconditional transfers. The sweet spot, conditions that set a clear direction while leaving open the means of getting there, requires sophisticated, capable, and legitimized public institutions. This has direct implications for developing economies. The institutions required for effective conditionality (capable implementing agencies, independent verification systems, robust monitoring infrastructure, insulated technocracies) are the preconditions for industrial policy to be more than a subsidy program with branding.
What This Means for Policymakers
Together, these two publications offer a framework that is more practically useful than anything available even five years ago. For policymakers in developing economies, the synthesis demands five honest questions before any industrial policy is designed.
- What is the market failure, structural challenge, or public purpose goal that justifies intervention? Directionality must precede instrumentation.
- Which tools are actually feasible given the market size, fiscal space, and government bandwidth? The World Bank's typology is a useful filter. Don't reach for instruments you cannot operate.
- What specific firm behaviors are we trying to induce — and what conditions, attached to public support, will reliably produce them? Mazzucato and Rodrik's taxonomy is the design guide.
- How will risks and rewards be shared? If public money underwrites the downside, there must be a credible mechanism — royalties, equity stakes, repayment triggers — for public institutions to participate in the upside.
- Do we have the institutional capacity to monitor compliance, enforce conditions, and adapt iteratively? If not, building that capacity is the first industrial policy priority.
There is something quietly significant about this moment. The World Bank, once the institutional home of the Washington Consensus, is now publishing comprehensive guidance on how governments should actively shape their economies. Mazzucato and Rodrik, representing different traditions and institutional homes, are co-authoring a practical taxonomy of the conditionalities that can make those interventions serve the public good. However, the risk is that governments, encouraged by the new permissiveness toward industrial policy, launch ambitious programs without the conditionalities to ensure public benefit, and without the institutional infrastructure to enforce those conditions even if they existed. Subsidies flow. Firms capture rents. Outcomes disappoint. And the backlash has the potential to set the conversation back another decade.
That cycle is not inevitable. Both of these publications, read carefully and implemented seriously, contain the tools to break it. The question is whether policymakers and the international development community advising them will do the harder work of building institutions before they build policies, and of negotiating real conditions before they sign the checks. Industrial policy without conditionality is not an industrial strategy. It is wishful spending.
Vishnu Venugopalan was a Practitioner-in-Residence on the Reimagining the Economy Project in 2025.
Read the Sources
Both publications are open access and essential reading for anyone working in industrial policy and development economics.
