Trump’s incredibly dumb economic mistake

Imports don’t subtract from GDP, tariffs don’t reduce trade deficits

Donald Trump wants to impose a 10 percent tax (though sometimes he says 20) on all imported goods.

This is a bad idea that does not make sense. Kamala Harris’s team has taken to characterizing it as a de facto national sales tax, but that is much too generous an estimation.

A national sales tax (or a value-added tax) would raise the same amount of revenue with a much lower tax rate. People wouldn’t be happy about paying higher taxes, but the impact on the national economy would be pretty minor, and arguably beneficial because the budget deficit and interest rates would be lower. Where Harris is correct is that all the downsides of a modest national sales tax (things would be more expensive) also accrue to a 20 percent import tax. But because the 20 percent tax is such a high rate levied on such a narrow base, it’s also a huge economic distortion.

Tons of labor and capital will be directed away from its best use, the overall productivity of the economy will decline, and wages will fall.

If you talk to Trump supporters about this, you hear a range of responses:

1. One class of prominent Trump advocates, including Rich Guy Posters like Elon Musk and Bill Ackman, just completely ignores the issue. They talk, constantly, about the election and about why they want Donald Trump to win the election, but they have nothing to say about his main economic policy proposal.

2. You also hear from people who just assert that while this idea is kind of crazy, Trump won’t actually do it. They point out that when he was president, he raised tariffs, but did so in a much more measured way than he is currently proposing.

3. And finally, you have people like Oren Cass in the Atlantic pretending that Trump has some kind of sophisticated industrial policy strategy to improve America’s manufacturing base and get one over on China.

The problem with Cass-ism is that it’s totally unresponsive to what Trump is actually proposing. A tax on Mexican fruit or Canadian lumber doesn’t boost American manufacturing. Those aren’t manufactured goods! Steel and aluminum are manufactured goods, but they are also things that manufactured goods are made out of.

A country can absolutely boost its manufacturing sector by deviating from dogmatic free trade consensus. For example, while most passenger airplanes flown in the United States are made in the United States, some of them are made in Canada (by Bombardier) or Brazil (by Embraer) or in Europe (by Airbus). I don’t think imposing high taxes on imported airplanes would be a good idea because it would reduce airline competition and raise prices. But it would almost certainly lead to more airplanes being manufactured in the United States. Taxing everything simultaneously, though — including airplane components, raw materials that are used to make airplane components, and primary commodities like food that people need more than airplanes — does not boost airplane manufacturing. It simply makes the United States an undesirable place to locate an airplane factory.

So why does Trump propose this? The most generous interpretation is number two in our list above — he’s just bullshitting because he thinks it sounds good and he won’t actually do it.

That might be correct. Trump is a huge liar, and he’s susceptible to both bribery and flattery, and it’s certainly possible that he just won’t do it. That said, what I’ve heard from sensible people who worked mid-level jobs on economic policy for Trump, and who are proud of TCJA and of the Republican Party approach to energy issues, and who think Democrats have bad opinions on guns and trans rights and crime, is that Trump just has incorrect, stupid beliefs about trade. He’s a mercantilist who believes that countries get rich by accumulating money, so he believes it’s good to export (foreigners give you money) and bad to import (you give foreigners money), and tariffs are a perfect tax because they raise revenue while deterring people from engaging in the bad activity of giving money to foreigners.

Trump’s accounting identity fallacy

One reason that Trump’s ideas about this are dangerous and that the former officials I’ve spoken to are less sanguine than Trump’s most online supporters, is that he has cultivated a circle of people who agree with him.

The highest-ranking mercantilists in the Trump administration were Commerce Secretary Wilbur Ross and economic advisor Pete Navarro. And back in 2016, they put together a pseudo-sophisticated account of why crude efforts to reduce imports would be economically beneficial, based on the idea that, “When net exports are negative, that is, when a country runs a trade deficit by importing more than it exports, this subtracts from growth.” So, they explain, fewer imports means more growth:

To score the benefits of eliminating trade deficit drag, we don’t need any complex computer model. We simply add up most (if not all) of the tax revenues and capital expenditures that would be gained if the trade deficit were eliminated. We have modeled only the impacts of implicit profits and wages, not any other economic aspect of the increased activity.

Trump proposes eliminating America’s $500 billion trade deficit through a combination of increased exports and reduced imports. Again assuming labor is 44 percent of GDP, eliminating the deficit would result in $220 billion of additional wages. This additional wage income would be taxed at an effective rate of 28 percent (including trust taxes), yielding additional tax revenues of $61.6 billion.

There are plenty of people on the internet who believe this based on the dangerous habit of reading one page of an economics textbook.

They think Trump is correct that imports reduce GDP and that reducing imports will therefore boost GDP, because they have read the formula which says that GDP equals Consumption plus Investment plus Government Purchases plus Net Exports. Net exports equal exports minus imports. Therefore, the lower your imports, the higher your GDP.

This is not correct, and shows why it’s useful to either take an economics class or else read the entire chapter in which this formula appears.

I’m going to try to explain the problem a few different ways, but the core point is that the Net Exports term is in the equation to avoid double-counting. If something is imported, then it was imported by someone — it was either bought by households so it’s included in C, or it was bought by businesses so it’s included in I, or it was bought by the government so it’s included in G.

As a math exercise, you could write the equation as Domestically Produced Consumption + Domestically Produced Investment + Domestically Produced Government Purchases + Exports and just leave the imports out altogether. It’s not normally done that way because government statisticians can’t use that as a counting procedure. The equation explains how the GDP data is assembled, which mechanically involves adding consumption data to investment data to government purchase data to net export data. That’s how they add it up; it’s not a theory of economic growth.

Accounting identities aren’t causal theories

If you’re a conservative Trump fan who thinks I’m full of shit, consider this: Can a poor country make itself richer just by increasing government spending?

After all, there are two ways that the government can spend money. One is it can increase transfer payments to households, which will boost household consumption. The other is that it can boost government purchases — build new office buildings, hire more public employees, construct a road.

I think basically everyone would acknowledge that increasing government purchases sometimes has economic benefits (if the government builds something useful) and sometimes does not (if they waste the money). One of the major disagreements between the left and right is over the likelihood that government purchases will be useful rather than wasteful. But what absolutely nobody thinks — certainly not conservatives — is that arbitrarily large increases in government spending generate arbitrarily large amounts of economic growth. A lot of people think this can work if the economy is in a steep recession; that’s Keynesian stimulus. But endless increases in government purchases are not a formula for an upward spiral of prosperity, and I don’t think anyone truly believes that they are.

A poor country like Guatemala can’t become rich simply by spending more and more. Are there smart, targeted investment Guatemala could make to become more prosperous? I bet there are. And I’d also bet there are dumb, wasteful things the Guatemalan state is spending money on that should be cut. But either way, to make the correct decisions, the government needs a detailed analysis of the situation, not just the observation that G occurs in the national income statistics.

Because the question this formula is answering is not, “How do you make your country rich?” but, “What statistical process do you use to know how rich the country is?”

Notably, there is more than one statistical process that you could use. And I think the world could be spared a lot of confusion if we used the other one instead. The way this other formula works is:

Gross Domestic Income = Wages + Profits + Taxes - Subsidies

I find this conceptually easier to understand. You’re looking at all the money that was earned (either as wages or business profits or something like rent or interest income) and then adjusting for net taxes. That’s the country’s total income. And the fun part is that because all the income that’s earned is earned by selling some kind of good or service, the market value of all income (GDI) equals the market value of all production (GDP), and there you are, measuring the economy with no import term at all.

The US government does, in fact, calculate both GDP and GDI. And the funny thing is the numbers don’t quite equal each other, even though they’re supposed to be equal, by definition. That’s a valuable reminder that we’re talking about a statistical procedure. The real world is messier than the abstract concepts, and two different ways of counting the same thing give you slightly different answers because of measurement error. But, again, this is a counting procedure, not a theory of economic growth. Taxes are added to GDI, but that doesn’t mean that raising taxes mechanically increases national income. What’s true is that in a growing economy, you generally expect wages and profits and tax revenue to all rise.

Financial flows need to add up

The other thing that most people, including lots of politicians, get wrong about this is that all the different cross-border financial flows need to add up.

America has a lot of successful high-tech companies. People all around the world use Microsoft Office and Gmail and Netflix. And even though iPhones are not manufactured in the United States, the majority of the price of an iPhone is Apple’s software and design work. People outside of the US also watch a lot of American movies and television shows and listen to a lot of American music. Every sale of a Taylor Swift album to someone in Italy causes a financial flow into the United States.

The United States is also considered a good place to invest.

Our government is very creditworthy if you want to make a safe, low-yield investment. There are a lot of individual American companies people want to invest in, but beyond that if you’re just looking to buy an index fund, we have a good overall stock market. There’s always plenty of buyers and plenty of sellers. The SEC is well-established and knows what it’s doing. Our companies are not going to be nationalized. There isn’t going to be a Communist revolution. America also has much faster population growth than Europe or Japan or Korea, so as a kind of broad default, investing in generic stuff like houses and commercial buildings is appealing. All the people in other countries investing in the United States create financial inflows.

You look at those services inflows and you look at the investment inflows, and the money has to go somewhere. Where it goes, right now, is that the United States imports more goods than it exports.

If you have specific supply chain concerns like, “We shouldn’t be getting all of our batteries from China,” it’s perfectly reasonable to try to address that through trade policy. Some of that should be tariffs or other moves to block Chinese products. Some of it needs to be regulatory reform to make it easier to build the things we want in America. But a lot of it should be freer trade with friendly countries, because I don’t think it makes a ton of sense to pursue a lower overall trade deficit as a policy goal.

To the extent that you disagree, though, the question is where do you want the money to go? Do we want to make American services exports less successful? Do we want to make the United States a less desirable location for foreigners to invest? The only halfway plausible answer would be that the federal government could run a persistent budget surplus. I don’t think anyone has a real vision for making that happen, but it’s what you’d need to do to make the whole thing work. Which returns us to our de facto national sales tax. It’s not just that an actual economy-wide consumption tax would be less distortionary than 20 percent tariffs. If you did a big tax hike like that to cut the budget deficit, you might actually end up with a lower trade deficit. You also might not — the world is complicated — but it would at least tend to push in that direction.

I’m not saying this is a good idea, but unlike what Trump is offering, it at least makes some kind of sense. And I’m begging those on the right who want to do something to promote American manufacturing or reduce the trade deficit to start saying things that make sense rather than sanewashing Trump’s ideas, which just reflect a widely held misunderstanding and not a real policy perspective at all.