Et Tu, World Bank? Industrial Policy on the International Scene

The World Bank recently published a 276-page report supporting the idea that industrial policy belongs “in the national policy toolkit of all countries.” This is a significant reversal for an institution that spent decades pushing developing nations toward fiscal discipline, open trade, and market liberalization. When the World Bank seems more interested in engaging with right- and left-wing populism than in promoting good economics, it tells you a lot about the era in which we live.

Industrial policy refers to government officials channeling resources to particular industries that the market would not. Arguments like national security or protecting “strategic” industries from competitors are often used to justify the policy. Whatever one thinks of these excuses, industrial policy is funded by taxpayers when the chosen instrument is subsidies, funded by consumers when the tool is tariffs, and always funded by the other domestic firms quietly crowded out as capital flows toward their politically favored competitors.

Every dollar directed by bureaucratic decree is a dollar that’s no longer directed by people spending their money on what most deserves it. Which, of course, is what makes markets work.

To be clear, the World Bank’s reversal wasn’t because a new generation of economists finally cracked open the historical record and discovered that state-led industrialization works. It’s because the World Bank’s most powerful shareholders, the United States and Western Europe, turned toward openly and aggressively practicing industrial policy.

With a cascade of green industrial subsidies during the Biden and Obama administrations, and protectionist tariffs and “golden shares” under the Trump administration, it became impossible to lecture developing countries about the dangers of letting governments pick winning businesses. In other words, the intellectual reversal followed the political reversal, not the other way around.

The World Bank’s report thus exists as a manual for governments that are going to do industrial policy regardless of what anyone tells them. It starts with the acknowledgment that all 183 countries surveyed boost at least one industry. But it stops arguing about whether industrial policy is legitimate and instead tries to diagnose which tools governments are capable of using without doing more harm than good. Mapping 15 different policy instruments along a spectrum from simple and low-risk to complex and demanding, the report warns governments repeatedly against blunt instruments that are politically easy but economically costly and urges governments to listen.

They won’t listen, and here’s why.

The report acknowledges that governments regularly botch industrial policy, yet it expresses hope that rising global education levels are giving more countries the human capital to make certain tools work. For example, a software tax exemption in Romania succeeded partly because a critical mass of people was now capable of becoming software engineers. Fair enough.

But while education raises the ceiling on what is theoretically achievable, it does nothing by itself to change a government’s incentives. The obstacle has never primarily been a shortage of capable technocrats or populations. The real hindrances are well documented, structural and bipartisan.

The first obstacle is what economists call the “knowledge problem.” As the Cato Institute’s Scott Lincicome notes, centralized attempts to identify critical technologies repeatedly fail because governments cannot predict which will end up being most valuable or how markets will develop. In the 1990s, governments picked the right industries — semiconductors and supercomputers — but the wrong products and companies. No amount of educational attainment by bureaucracies or workers solves this. Only markets aggregate untold amounts of economic knowledge through rough, supply-and-demand responsive prices and voluntary exchange.

The report never grapples with this, suggesting tools like industrial parks aimed at coordination failures and skills-development programs aimed at under-investments in human capital. Someone must still decide where the park goes and which skills get funded for which sectors. These are predictions about what the economy will need, made by the same officials facing the same information constraints as any other planner. They are dressed in more sophisticated language than a tariff but are no less vulnerable to being wrong.

The second obstacle is politics. Educated people, bureaucrats and CEOs operate under and inside governments where industries lobby, ministers have constituencies and failing programs are far easier to throw more money at than to kill. The World Bank’s report concedes as much in calling the bluntest instruments “notoriously difficult to unwind.” But that’s not a technical or educational problem; it’s a political one.

Even in a country as educated as the U.S., a steel industry that has enjoyed decades of political protection does not quietly accept a withdrawn subsidy. It assumes the role of political actor, pressuring politicians for more. As protectionism dulls the industry’s genuine competitiveness, politics matter even more. This is bad even for nations rich enough to shoulder the cost.

The World Bank spent 276 pages telling governments how to do more of what governments most want to do but rarely do well. For developing nations, it’s like receiving a lifejacket that works better for people who already know how to swim and can afford to paddle around.