The economics theory that could have saved the Trump presidency

Sometimes it's good to listen to the economists.

Source: Diamond & Mirrlees (1971)

Americans are not very happy with Donald Trump’s second term in office. Trump’s approval has trended downward since he returned to power, and has recently fallen again. Here are some numbers from Nate Silver:

Source: Nate Silver

Even Fox News admits how bad it’s getting:

Unhappy with the economy. Pain with prices. Unsure about Trump administration policies. It adds up to high disapproval among the president’s loyal constituencies…

Some 76% of voters view the economy negatively. That’s worse than the 67% who felt that way in July and the 70% who said the same at the end of former President Biden’s term…Large numbers, overall and among Republicans, say their costs for groceries, utilities, healthcare and housing have gone up this year…Voters blame the president. About twice as many say President Donald Trump, rather than Biden, is responsible for the current economy. And three times as many say Trump’s economic policies have hurt them (they said the same about Biden’s last year). Plus, approval of how Trump is handling the economy hit a new low, and disapproval of his overall job performance hit record highs among core supporters…

Trump’s job performance drew career-high disapproval among men, White voters and those without a college degree…Among all voters, 41% approve of the job Trump is doing, while 58% disapprove…For comparison, Biden’s marks were a bit better at the same point in his presidency: 44% approved and 54% disapproved in November 2021.

As Fox points out, it’s the economy, rather than immigration policy or culture wars, that’s driving these results. Despite employment, growth and inflation numbers that don’t seem too bad overall, Americans are deeply unhappy with their economy. Preliminary numbers for November’s consumer sentiment show it falling back to the low point it hit in 2022, during the height of the post-pandemic inflation:

Source: UMich

And as Fox notes, voters overwhelmingly blame President Trump, rather than Biden, for the economy. Even 42% of Republicans blame Trump rather than Biden, which must be a galling thing to tell a pollster on the phone. Analysis by David Shor finds that Republicans’ trust advantage is evaporating on issues like the cost of living, the budget deficit, and the economy in general.

Exactly why Americans hate their economy right now is a tough question. Sometime during the Biden administration, we saw economic sentiment decouple from the macroeconomic numbers that traditionally correlated with sentiment — in other words, we saw the beginning of the “vibecession”. Now after a modest recovery in 2024, we’re seeing another vibecession under Trump. What’s going on? Is it mortgage rates? Anxiety about AI? Are people displacing their concerns about social unrest onto their perceptions of economic conditions?

One possibility is that Americans are expressing their unhappiness about economic policy, rather than economic results. Both Trump’s approval and consumer sentiment fell abruptly during the recent government shutdown, and they also hit a previous low in May after Trump’s “Liberation Day” tariffs were announced. So low consumer sentiment, and low Trump approval ratings, might be Americans’ way of expressing unhappiness about what Trump is trying to do to the economy.

People must have an intuitive sense that the AI boom is the main thing propping up the macroeconomy right now, and that this could end at any moment. And they probably realize that the AI boom is having to fight against the headwinds created by Trump’s tariffs. They can also see that tariffs are causing localized harm to parts of the U.S. economy right now.

The part of the economy being hurt the most is manufacturing — exactly the sector that Trump has long pledged to help. Even as employment holds up in service jobs like health care and education, employment in goods-producing industries has plunged since “Liberation Day”:

Source: Justin Wolfers

In fact it’s not just manufacturing that’s hurting. Construction and transportation/warehousing jobs, which were booming in Biden’s final year in office, have basically collapsed under Trump. A lot of that construction was factories, which boomed more under Biden than at any time since the 1960s. Under Trump, the factory construction boom has begun to deflate.

But no sector is hurting worse than manufacturing, which continues to shed jobs at a rapid rate. Almost every type of manufacturing is doing badly, but the auto industry has swung from expansion under Biden to contraction under Trump:

Source: Joey Politano

What’s going on here? The obvious answer is “tariffs”. Here’s Reuters from a couple of weeks ago:

U.S. manufacturing contracted for an eighth straight month in October as new orders remained subdued, and suppliers were taking longer to deliver materials to factories against the backdrop of tariffs on imported goods…Backlog orders remained subdued as did export orders…Production was weak after briefly rebounding in September. Manufacturers have cited tariffs as a major constraint…

Tariffs are gumming up supply chains, resulting in longer delivery times to factories. The ISM survey’s supplier deliveries index increased to 54.2 from 52.6 in September. A reading above 50 indicates slower deliveries…Factories continued to pay more for inputs[.]

This story is not a one-off. In fact, we have seen a steady drumbeat of stories about American manufacturing’s tariff-induced woes since May. In September, Moody’s Analytics assessed that manufacturing was experiencing recession-like conditions, thanks to tariffs. The Institute for Supply Management, which interviews manufacturers, has been receiving an avalanche of tariff-related complaints, along with consistent pessimism about business conditions, for months now.

Things are not as catastrophic as they might have been. Resource prices have fallen worldwide, which has helped cancel out some of the tariffs’ impact. And Trump has backed off of some of his tariffs, especially on China, while granting a byzantine maze of exemptions and deferrals. The effective tariff rate on imports has risen, but only to 10.5% so far — considerably less than the headline rates that Trump has been throwing around on the news.

But although tariffs are hurting less than they would have if Trump had stuck to his guns, they’re certainly hurting to some degree. It’s entirely reasonable for the American people to be unhappy about their government intentionally hurting the economy, even if it doesn’t end up being as bad as promised.

It didn’t have to be this way. If Trump had listened to economists, he might have known that tariffs don’t work the way you might assume from watching CNN in the early 1990s — which, as far as I can tell, is where Trump got most of his ideas about the way the world works. If Trump had listened to economists, he might have understood why tariffs hurt manufacturing.

The reason is that tariffs include taxes on intermediate goods, which make production less efficient.

I’ve talked about this a number of times in the past, and I’ve shown some empirical evidence that tariffs on intermediate goods really do hurt U.S. manufacturing. But I’ve never really talked about the theory behind this idea. I really should talk about it, because if Trump and his people had understood this concept, they might have avoided a lot of pain and a lot of mistakes.

The theory originally comes from a 1971 paper by Peter Diamond and James Mirrlees. Diamond and Mirrlees realize that the government needs to tax the economy in order to produce various stuff (highways, education, research, etc.), and to redistribute income via the welfare state. Most types of taxes tend to distort the economy.¹ But Diamond and Mirrlees showed that if the government can tax everything — cars, pizza, back massages, labor, and so on — at different rates, you can levy taxes without distorting economic production.

Now, that’s not that useful of a result. In reality, you can’t actually put different tax rates on every different kind of good or service. But the real value of the Diamond-Mirrlees result is that it shows what kind of taxes you don’t want to use: taxes on intermediate goods. They show that taxing intermediate goods is always worse than taxing either final goods — i.e., stuff consumers buy, like cars and pizza — or “factors of production” (i.e., labor, capital, land, etc.). You never want to tax stuff like steel, or auto parts, or computer chips that companies buy in order to produce other stuff.

If you like, here are some slides from Todd Lensman that explain the math of this result in a simplified form.

So why don’t you want to tax intermediate goods? Because you want to make as much stuff as you can make before you start redistributing it. If the purpose of taxes is to redistribute the economic pie,² you want to redistribute as big of a pie as you possibly can. Taxes on things like steel, auto parts, and computer chips cause the economy to make fewer cars, houses, computers, etc. So when you collect taxes and then give people money to spend, there are fewer cars, houses, and computers for them to buy with the money you give them. You should have just let capitalism work its magic and make as much stuff as possible, and then worried about how to redistribute.

This is a very powerful and deep result. Diamond and Mirrlees made some simplifying assumptions in order to make the math easier, but other theorists came in later and did the harder math, and they showed that the basic result — no taxes on intermediate goods — holds for a pretty wide range of assumptions.

That’s a very useful real-world result! In fact, real-world tax systems in rich countries mostly stick to the Diamond-Mirrlees principle. Income and payroll taxes are taxes on “factor inputs”, so they’re OK. Corporate taxes allow you to deduct business expenses, so they’re also OK — you’re not taxing the cost of the intermediate goods that businesses buy.

U.S. sales taxes are actually bad, because they violate the Diamond-Mirrlees principle — lots of sales taxes are levied on B2B transactions. Europe does this much better — their value-added tax (VAT) is basically a sales tax that doesn’t get charged on the things businesses buy. America would be a bit richer if we switched from sales taxes to VAT, but we won’t, because sales taxes are used by our state and local governments, while a VAT would have to be nationally administered.

Anyway, Trump’s tariffs absolutely violate the Diamond-Mirrlees principle. When most people think about imports, they think about cheap stuff you buy on the shelf at Wal-Mart. Diamond and Mirrlees would be fine with taxing that stuff — the things on the shelf at Wal-Mart are “final goods”. But almost half of what America imports from overseas is intermediate goods. Here are numbers from 2019:

Source: SF Fed

Trump’s tariffs apply to all of those intermediate goods. That’s why carmakers are having trouble making cars right now. And it’s why even if Trump does mail tariff rebate checks to every American, the number of cars they’ll be able to buy with those checks will be fewer than they could have bought before the tariffs.

Because neither Trump nor any of his people understood the basic insight of Diamond-Mirrlees (1971), they are trying to redistribute a pie that they’ve already shrunk. And the American people don’t appear to be happy about it.

Are there ever situations where you’d want to tax intermediate goods? Yes. Like any economic theory, if you break enough of the assumptions, the basic result no longer holds. Costinot and Werning (2022) show that if your only taxes are income taxes and tariffs, you should use tariffs to reduce inequality. Basically, if imports from China hit a few kinds of American workers very hard, and you have no way to specifically compensate those folks,³ then you should have some tariffs on China, in order to protect those few workers. But even then, Costinot and Werning show that the optimal tariff will be very small — between 0.02% and 0.12%, compared with the 10% that Trump has enacted so far. The world isn’t exactly like Diamond-Mirrlees, but it’s pretty close.

So anyway, tariffs on intermediate goods are bad. Economists knew this, and shouted it from the rooftops. But the Trump administration prides itself on not listening to economists. JD Vance has declared that “the economics profession doesn’t fully understand tariffs”, and hardly a day goes by when Trump-aligned intellectuals like Oren Cass don’t sneer at the economics field.

And yet this willful ignorance comes with real political costs. No, economists don’t know everything about how the economy works. Maybe they don’t even know most things. But they do know some things, and one of those things is that taxing intermediate goods hurts the economy.

If Trump’s people had allowed themselves to understand that fact — if they had listened to the economists — Trump’s approval ratings might not be nearly so low as they are.

And if you’re a progressive, the temptation to laugh at Trump over all of this will naturally be very strong. But the right lesson here isn’t that “Trump is dumb” (though that is probably true). The right lesson here is that although there are lots of things they don’t know, and although they don’t get everything right, and although they’re often overconfident, economists are worth listening to, even when they don’t tell you what you want to hear. That’s a lesson that Biden and his people should have heeded before they unleashed a stimulus plan that macroeconomists predicted would exacerbate inflation. And it’s a lesson that will come in handy the next time Democrats are in charge of the economy.

---

1. Except for land taxes and “lump-sum” taxes. Lump-sum taxes are where the government says “OK, every citizen give me $100”. Obviously that’s not going to fly in real life. Land taxes actually do work in real life, although they can logistically be tough to implement, due to the difficulty of distinguishing between the value of land and the value of all the stuff that’s built on top of the land.

2. Diamond and Mirrlees also allow for another purpose of government: to produce useful stuff, like infrastructure and education. But here, their solution is basically just “run the government production exactly like a private company would run it” — essentially, make some SOEs that minimize costs like a private company would. Easier said than done, right? But anyway, that doesn’t really affect the “intermediate goods” result.

3. From a Diamond-Mirrlees standpoint, the best way to compensate these folks would probably be to subsidize the industries that are subject to the most intense Chinese competition, rather than to send money to workers who get displaced.