The Latest Thinking on Biden’s Biggest Economic Idea
Industrial policy was a punching bag for economists for decades. Governments couldn’t possibly know enough to pick winners and losers, economists averred. And even if policymakers did know, special interests would capture the process and steer taxpayer money to themselves.
You can still find economists saying things like that, because the concerns are genuine. Industrial policy really can be costly and counterproductive.
But the disdain of economists never managed to stop governments from conducting industrial policy. And there’s mounting evidence that the right kind can work as intended — and has. That’s good news for the Biden administration, which has gone all in on industrial policy with the American Rescue Plan, the Infrastructure Investment and Jobs Act, the Inflation Reduction Act and the CHIPS and Science Act.
A new review of the economic literature points to how industrial policy can succeed as well as to how it can fail. “The New Economics of Industrial Policy,” issued as a working paper last month, is a draft of an article for the Annual Review of Economics. It’s by Reka Juhasz of the University of British Columbia, Nathan J. Lane of the University of Oxford and Dani Rodrik of Harvard’s Kennedy School of Government.
Rodrik, who is the best known of the three, was positively inclined toward industrial policy even when few other economists were. In an interview, he conceded that some skepticism about his impartiality as a reviewer of the literature is therefore “healthy.” But he told me the studies speak for themselves. “The studies that are empirically more careful, post-1990 — the least you can say about them, if you’re still anti-industrial-policy, is that they do not present a categorically negative case for industrial policy in the way that the earlier generation did,” he said.
There are three justifications for industrial policy, each based on correcting a failure of the free market. One is that companies left to their own devices invest less than is socially optimal because they capture only some of the benefits; targeted measures to increase their investment can benefit all. Another justification is that private companies don’t manage to coordinate their efforts in a way that is good for all; the government can be the coordinator. Third, private companies require certain inputs that only government can reasonably supply, such as a highway to a port.
Up against these three positives are the two negatives, namely the risks of government incompetence and regulatory capture.
Industrial policy can appear to be a failure if all you do is naïvely compare the performance of a sector with how much government support it receives. The problem with that simple correlation is that some sectors are targeted for help precisely because market failures are more of a problem for them. They might have done even worse without government intervention.
Tariffs are the best-known form of industrial policy but possibly the worst because they raise prices for consumers. Fixing coordination failures is a lighter touch that can work nicely. “Since government attention is a scarce good, public-private collaboration focused on alleviating constraints faced by specific sectors or groups of firms, such as deliberation councils or business-government round tables, also counts as industrial policy,” the authors of the literature review wrote.
Governments — most, anyway — seem to have realized that tariffs don’t work as advertised. Tariffs account for just 1.3 percent of industrial policy interventions globally and 3 percent in the bottom two-fifths of countries by income, according to the literature review.
Critics of industrial policy worry about companies and government winding up in bed with each other, but done right, “embeddedness” can actually be a highly successful form of industrial policy, the authors wrote. They cited the sociologist Peter Evans, who wrote in 1995 that social ties between executives and officials provided “institutionalized channels for the continual negotiation and renegotiation of goals and policies.”
The Defense Advanced Research Projects Agency, which played a key role in the development of the internet, GPS and graphical user interfaces, among other technologies, uses something like Evans’s “embedded autonomy” in its dealings with the private sector, the authors wrote.
In the United States, the gearing of industry for military production in World War II and the Apollo lunar mission were two signal successes for what we would today call industrial policy, according to recent studies cited in the literature review. They caused lasting increases in manufacturing output.
One caveat: Not everybody wins. The long-run benefits from the federal government’s war-related funding of R. and D. were felt by only 5 percent of technology clusters, namely those in counties that were most innovative in 1930, even before they received the government contracts, according to a paper the authors cited, by Daniel Gross of Duke’s Fuqua School of Business and Bhaven Sampat of Columbia. “This is a sobering insight for initiatives that envisage similar public R. and D. investment in less auspicious locations today,” the literature review said.
That “sobering insight” points to a bigger potential problem with industrial policy, which is that its objectives can conflict. President Biden wants to reduce greenhouse gas emissions and create more high-paying factory jobs. But the Commerce Department’s recent levying of high tariffs on solar panels from Southeast Asia, which are intended to protect U.S. manufacturers, is likely to slow America’s green transition.
I asked Rodrik about those tariffs. “That’s a valid objection,” he said. “We have to be careful about not overpromising.”
On the whole, Rodrik said, he thinks the Biden administration’s industrial policy is on the right track. Industrial policy can happen by accident or with intent, he said. “It’s got to be better that they’re doing it in a purposeful and systematic way.”"