Michael Pettis's Unpersuasive Case for Industry Subsidies

In a Wall Street Journal article titled “Are Domestic Manufacturing Jobs Worth Protecting?” (May 1, 2025 print), Michael Pettis answers yes, if the reason is that the job loss has come about, not due to increases in productivity but due to other countries’ governments subsidies to manufacturing.

What does Pettis think is the reason that those governments subsidize manufacturing? He writes:

These countries [he means “these countries’ governments] believe that the gains from productivity improvements in manufacturing will spill over into the broader economy.

Pettis may be right about their beliefs. He seems to making a case that there are positive externalities from manufacturing. He doesn’t, though, get around to naming any.

Pettis claims that letting “other countries take the lead in manufacturing harms the American economy.” How so? He writes:

The U.S. accommodates other countries’ successful industrial policies by absorbing their negative consequences—namely, by absorbing global savings imbalances, running trade deficits and offshoring manufacturing.

He’s way off. Let’s look at each of the three. “Absorbing global savings imbalances” means simply that people in other countries invest more here (not just savings in the form of buying U.S. government bonds but also actual investments in land, plant and equipment, and stocks) than we invest there. He doesn’t say why that’s bad. Second, he thinks that running trade deficits is bad. But trade deficits are the mirror image of capital surpluses, which is the more-neutral term for “global savings imbalances.” So he’s simply restating his first point. Third, if a government doesn’t subsidize manufacturing, it will have less manufacturing, meaning that some (not all and not close to all) manufacturing will be offshore. The question is why that’s bad. You can’t just assume it, especially in an op/ed where you’re trying to make the case for subsidies.

Pettis states that the U.S. government’s failure to play tit for tat with its own subsidies to manufacturing has bad consequences:

America’s decision not to encourage manufacturing has hollowed out industrial ecosystems and left key supply chains vulnerable. It has slowed productivity growth and weakened wage gains, especially for the bottom half of wage earners. And it has hindered America’s ability to respond flexibly to global shocks, from pandemics to geopolitical crises.

There’s so much to unpack in those three sentences.

The first sentence is incredibly vague. Which “ecosystems” have been hollowed out? Which key supply chains are vulnerable? He doesn’t say.

Pettis’s second sentence is almost certainly wrong. If we avoid subsidizing particular sectors, then we have a more efficient economy, which means that overall productivity is higher than otherwise.

His third sentence is too vague also. How has not subsidizing industries hindered America’s ability to respond flexibly to global shocks? A few years ago, I gave a talk to an Osher Lifelong Learning Institute (OLLI) audience in which I made the case for free trade. A member of the audience argued for having government favor domestic manufacturing. He said that if we had had more domestic production of masks during the Covid pandemic, we would not have been as vulnerable as we were because, at the time, we depended on imports from China. I agreed with him. I then asked him whether he had advocated more domestic production of masks before Covid came along. He admitted that he hadn’t. He saw my point. How does a government decide which industries to favor if it’s worried about global shocks. It can’t know what the shocks will be. It’s so easy to look back and say that you wish you had done X, when in fact, you would have had to choose between X and all the other letters of the alphabet, not knowing which one will be relevant.

Moreover, Pettis doesn't the address the combined donkey and elephant in the room: the fact that politicians in the Democratic and Republican parties will try to push for their favored industries to be subsidized, independent of the merits of the issue.

Is there anything to like in Pettis’s op/ed? Yes. At least he admits that losing manufacturing jobs due to higher productivity is a good thing. It would have been nice, though, if he had pointed out that over 80 percent of the loss in manufacturing jobs is due to increases in productivity. It would have been even nicer, but it might have undercut Pettis’s narrative, if he had pointed out that U.S. manufacturing output in the last quarter of 2024 was only 6.6 percent below its all-time high.