Once again: Imports do not subtract from GDP

I've explained it before, and I'll keep explaining it until the world understands.

Photo by SJ on Unsplash

Tariffs feature prominently in this year’s election debate. Although the Biden administration has put some pretty steep tariffs on some Chinese goods, Trump wants to go much further — he wants to put a 20% tariff on every single thing the U.S. imports, from any country. Trump seems to be a deep believer in the power of tariffs to solve almost any economic problem.

But why does Trump believe this? Matt Yglesias has been talking to some of Trump’s policy people from his first administration, and he has an answer. It’s because a whole lot of people think that imports subtract from GDP:

This is a very plausible explanation. Peter Navarro, one of Trump’s trade advisors, certainly believesthat imports subtract from GDP. But so do a whole lot of other people — including many people who don’t even support Trump.

So this is a good time for me to write a post explaining why this is a mistake. Despite what you may think you learned in econ 101, imports do not subtract from GDP — in fact, they’re not counted in GDP at all.

The decision to import more stuff can affect GDP indirectly. But that’s not the same. And whether importing more stuff causes GDP to go up or down depends on what the imports are used for, and a bunch of other factors as well. In fact, there are many cases where blocking imports causes GDP to fall — including, probably, Trump’s tariffs during his first term.

In fact, this is not the first time I’ve written this post. I wrote an op-ed for Bloomberg explaining it back in 2016, and I wrote another post explaining it for this blog in 2022:

Imports do not subtract from GDP

Obviously not everyone in the world reads this blog, so I can’t exactly expect these posts to have changed popular understanding. Nevertheless, the return of Trump and his tariffs means it’s a good time to try to explain it once again. And instead of simply reposting my previous posts word for word, I want to try to hone my explanatory skills.

OK, so first let’s explain why the popular pseudo-econ factoid is wrong, and imports don’t subtract from GDP.

Understanding why imports don’t subtract from GDP

Often, there are cases where economists understand some principle of the economy (for example, comparative advantage or opportunity cost) that normal folks find it difficult to wrap their mind around. The relationship of imports to GDP is not one of those cases. Instead, it’s a case where economists decided to teach a concept in a confusing way, resulting in widespread popular misunderstanding.

Here’s how your Econ 101 textbook probably taught you to calculate GDP:

GDP = Consumption + Investment + Government purchases + Net exports

And since net exports are just exports minus imports, this means:

GDP = Consumption + Investment + Government purchases + Exports - Imports

Look! Right there, it says “- Imports”, right there in the dang formula! Plain as day. You’d have to be a total idiot not to understand that simple formula, right? And indeed, some people responded to Matt Yglesias’ tweet by calling him an idiot.

But Matt is not an idiot. And neither, I like to think, am I.

Let’s talk about what GDP is. GDP is the total value of everything produced in a country:

GDP = all the stuff we produce

Imports aren’t produced in the country, so they just don’t count in the formula above. And they aren’t alone. There are plenty of other important things in the Universe that have don’t get counted in GDP, simply because they have nothing to do with domestic economic production. The number of asteroid impacts in the Andromeda galaxy is probably important to someone, but it doesn’t count in U.S. GDP. The population of the beluga sturgeon in Kazakhstan is probably important to someone, but it doesn’t count in U.S. GDP. Imports don’t count in U.S. GDP because, like asteroid impacts in the Andromeda Galaxy and the population of beluga sturgeon in Kazakhstan, they don’t involve domestic economic production in the United States.

In fact we can divide GDP up a different way from the Econ 101 breakdown. Let’s divide it up according to all the categories of people who might ultimately use the stuff¹ we produce in the U.S.:

GDP = Capital goods we produce for companies + Consumer goods we produce for consumers + Stuff we produce for the government + Stuff we produce for foreigners

Again, imports are nowhere to be seen. But exports are in here! Exports are just all the stuff we produce for foreigners. So the formula is:

GDP = Capital goods we produce for companies + Capital goods we produce for consumers + Stuff we produce for the government + Exports

This is a perfectly good formula for GDP. But instead, here’s what economists do. They add imports to the first three categories, and then subtract them again at the end:

GDP =

(Capital goods we produce for companies + Capital goods we import for companies)

+ (Consumer goods we produce for consumers + Consumer goods we import for consumers)

+ (Stuff we produce for the government + Stuff we import for the government)

+ Exports - Capital goods we import for companies - Capital goods we import for consumers - Stuff we import for the government

This type of equation adds in three different types of imports, then subtracts them all again at the end. It’s mathematically equivalent to the formula above it, because if you add imports and then subtract them out again, you’ve just added 0. And adding 0 does nothing. Imports still don’t count in GDP in this equation.

OK, now let’s realize what the terms in the equation mean:

• (Capital goods we produce for companies + Capital goods we import for companies) is just Investment.

• (Consumer goods we produce for consumers + Consumer goods we import for consumers) is just Consumption.

• (Stuff we produce for the government + Stuff we import for the government) is just Government purchases.

• Exports - Stuff we import for companies - Stuff we import for consumers - Stuff we import for the government is just Exports - Imports.

So the equation is now:

GDP = Consumption + Investment + Government purchases + Exports - Imports

This is just our good old Econ 101 equation. It looks like imports are being subtracted from GDP, but now you (hopefully realize) that this is because imports are also being added to consumption, investment, and government purchases! Consumption, Investment, and Government purchases include imports, so we subtract out imports at the end so that the total effect of imports on GDP is zero.

(If you still don’t believe me, go to my previous post and read the example of a lazy useless moon base. The moon base produces nothing at all, so its GDP is always exactly zero. Even if it imports stuff from Earth and runs a trade deficit, it’s GDP is still zero.)

So why did economists decide to write things down in this confusing way — invisibly adding imports to C, I, and G, and then explicitly subtracting them again from NX? Because all of the things in the Econ 101 textbook version are easy to measure in the real world. It’s hard to know exactly how much of the value of what we consume and invest was imported. So instead, our government agencies don’t try. They just subtract out imports at the end of the equation, so that the imports hidden within C, I, and G get canceled out.

Matt Yglesias understands this. The people calling him an idiot on Twitter do not understand this. And Peter Navarro, and apparently some of the other people who have worked for Trump, also do not understand this.

Why don’t they understand this? A bit of it is probably due to motivated reasoning — Trump has decreed that Tariffs Are Good, so people who support him must now come up with reasons to agree. But some of it probably comes from the fact that the economists’ confusing formulation — the “GDP = C + I + G + NX” equation — has become the basis for how some news outlets report quarterly growth numbers.

Here’s a quote from the New York Times in April 2022, which I complained about in my previous post:

The ballooning trade deficit, meanwhile, took more than three percentage points away from G.D.P. growth in the first quarter. Imports, which are subtracted from gross domestic product because they are produced abroad, have soared in recent months as U.S. consumers have kept spending. But exports, which add to G.D.P., have lagged in part because of weaker economic growth abroad. (emphasis mine)

The NYT is simply wrong. Imports are not subtracted from GDP, as we have seen. They simply don’t enter the GDP equation at all, positively or negatively. But it has become standard practice for economics reporters to just go down the list of C, I, G, and NX every quarter, reporting how much each of these “contributed” to growth. Every single quarter you’re talking about how imports lowered NX, while you go your whole career without ever once mentioning how imports raised C, I, and G by an equal and opposite amount. Out of sight, out of mind. And then the general public learns this way of thinking as well.

But there’s one more big reason why lots of people mistakenly think imports reduce GDP. It’s because imports can have indirect effects on productive economic activity. In some cases, this actually can reduce GDP! In other cases, it raises GDP.

Imports can affect GDP indirectly, but they might make it go up or down

Suppose you’re an American consumer. You see a shiny new electric car from BYD at an auto show. You think “Wow, that car is amazing…I need one of those!” So you buy it and take it home. The BYD car was made in China, so it’s an import. As I explained in the previous section, buying that car doesn’t subtract from GDP.

But if you hadn’t bought that BYD car, you would have bought an American-made car from GM instead. So now, because you didn’t buy a GM car, GM produces less stuff in America, and GDP goes down. Your purchase of the BYD car didn’t directly subtract from GDP, but it indirectly caused a chain of events that led to GDP decreasing.

This sort of case is probably what Trump and his people have in mind. If this were the only way imports worked, they’d be right — just slapping tariffs on foreign products would cause American consumers to buy American instead, and local production lines would start humming.

But now let’s consider some different cases. Suppose that Boeing needs to import some very high-quality steel from Arcelor in France in order to make a new model of airplane.² But suppose the U.S. puts tariffs on all foreign steel, so that Boeing can no longer afford to import it. Instead, Boeing looks around for an American steel supplier who can deliver what Arcelor could. Maybe there’s one small high-tech company that can deliver a little bit, but its prices are very high, and its production volume is small — and would take many years to ramp up.

Maybe you’re thinking that Boeing could launch a crash program of self-sufficiency — investing a ton of money in the one high-tech American steel startup that could meet its quality requirements, helping it to expand production. But that would take years, and it would be expensive, and it might not work. Instead, it’s a lot easier for Boeing to just cancel the new model of airplane. In this example, reductions in imports have indirectly made GDP go down, by stopping a U.S. manufacturer from being able to get critical parts and components.

In fact, this is probably what happened with Trump’s steel and aluminum tariffs. Lake and Liu (2022) show that Bush’s tariffs on steel a decade earlier harmed America’s steel-using industries more than it helped the American steel industry:

President Bush imposed safeguard tariffs on steel in early 2002…We find the tariffs did not boost local steel employment but substantially depressed local employment in steel-consuming industries for many years after Bush removed the tariffs. These large and persistent negative effects were concentrated in local labor markets that had low human capital or were strongly specialized in steel-consuming industries.

In other words, trying to protect America’s steel industry with tariffs ended up exacerbating the Rust Belt. Trump’s very similar tariffs probably had a very similar effect on U.S. manufacturing.

Tariffs on intermediate goods generally hurt GDP more than tariffs on final goods — steel tariffs are more likely to hurt regular Americans than tariffs on cars. That doesn’t mean tariffs on intermediate goods are always bad — for example, China supplies lots of intermediate goods to the U.S., and some of these might represent critical dependencies that we just can’t afford in the case of a major conflict. Tariffs are one way of reducing those vulnerabilities.

But in general, tariffs on intermediate goods should be avoided unless there’s a very compelling reason. Trump’s tariffs on U.S. allies during his first term, and his proposal to hit those allies even harder if he gets elected again, are certainly unhelpful from a geopolitical standpoint, but they also make America poorer.

In fact, the example of intermediate goods is not the only example of how a reduction in imports can indirectly hurt GDP. They can also hurt exports. Countries like Malaysia or South Korea need dollars to pay for U.S.-made products; they generally get those dollars by exporting things to the U.S. But if the U.S. shuts off imports from these countries, those countries will have more trouble getting foreign exchange — and U.S. exports will likely suffer as a result. And that’s not even bringing trade-war retaliation into the picture.

In other words, competition with U.S. companies — which a lot of people think about, and which Trump wants them to think about even more — is only one of several ways that imports indirectly affect GDP. The global economy is a highly complex machine of cross-border supply chains; we’re not the stand-alone economy we (sort of) were back in World War 2, and we can’t return to that level of self-sufficiency without making ourselves much poorer. Yes, supply chains that run through China are dangerous, and national security sometimes requires us to pay the cost of moving them out. But there is a cost, nonetheless.

I would like Americans to have a more nuanced view of the effects of imports and the effects of tariffs. To say that tariffs are never a good idea, and that we should always buy anything China wants to sell us at whatever price they offer, is naive and simplistic. But to imagine tariffs as a magic wand that boosts the U.S. economy is even more unrealistic. And tariffs on intermediate goods made by our allies, like the ones Trump is promising to impose, are a self-destructive policy that will make America both poorer and less secure.

---

1. The “stuff we produce” includes both goods and services. But because “goods and services” is cumbersome to type over and over again, and takes up a lot of space on the page, I’m going to just say “goods” instead. Please understand that this includes services.

2. The keen-eyed among you might be wondering if intermediate goods are counted as imports. Yes, they are. But since imports don’t count in GDP, intermediate goods imports also don’t count in GDP. The St. Louis Fed has a good explainer of how intermediate goods are accounted for in GDP calculations.