Why does a rise in imports lower gdp?

by on September 11, 2010 at 7:23 am in Economics | Permalink

FYI, a loyal MR reader and commentator, writes to me:

I would like to suggest a topic for your blog: the way imports are used in the GDP calculation. To me it makes no sense that we would decrease our GDP every time something is imported. What is the rationale here? That we are sending 'gold overseas'? I mean, if one really believes that trade works by exchanging items of same perceived value (I give you $50 and you give me a $50 product) how can we ever explain this GDP deduction? Isn't this another mercantilist heritage that proves how not so free our trade system is?

The rationale is a) that is simply how the number is calculated, and b) a nation's capacity to produce goods and services determines its long-run standard of living.  On b), you might argue that the world's capacity to produce goods and services is what matters in a globalized setting, which returns us to FYI's question.

The comments of Angus are on the mark ("People, Accounting identities do not imply causation!"):

As a matter of accounting, the arithmetic is correct. But the suggestion that the trade deficit CAUSED growth to be lower by some measurable amount is completely unproven and just plain wrong.

The argument implies that there somehow would have been perfect substitutes for all imported goods being produced domestically and available for sale at the same price. Thus, if we could just keep out those damn imports, growth and jobs would soar.

Yet, this is far from true on the face of it, let alone considering if we banned all imports, we'd have a pretty hard time making any exports and that might create a "drag" on growth too, no?

1 Ryan September 11, 2010 at 7:47 am

Maybe Keynes needed an equation that demonstrated that G increased Y, so he created one. Mises never understood what Y actually meant in this case, and he was a pretty smart guy. What is “the sum total of all monetary transactions within a nation’s borders?” Does such a number actually tell us something?

…Maybe those crazy Austrians are right, and the whole “equation” is a little crackers.

2 liberalarts September 11, 2010 at 8:26 am

I always explain it to my students this way:

GDP is a measure of total value of production for a country, but it is easier to measure total purchases via C + I + G, so we do that. We then have to add in exports and subtract out imports so that purchases equal the GDP measure of production, since exports are things that we produce but don’t buy and imports are things that we buy but didn’t produce.

Thus net exports are just included to make an imperfect approximation of GDP (domestic purchases) a more accurate of domestic production.

3 bob September 11, 2010 at 8:45 am

Imports don’t lower GDP. The accounting identity subtracts imports because C, I, and G include domestic expenditures on foreign goods, and GDP is supposed to be the value of domestic production.

If United buys an A380 from Airbus for $30 billion, both I and imports rise by $30 billion, and GDP doesn’t change.

4 spencer September 11, 2010 at 9:02 am

The national accounts estimate production, but they do not actually measure output.

Rather, they measure consumption and adjust that by the change in trade and inventories to indirectly estimate production.

Yes, it makes sense.We consume imports but we do not produce imports and that is what GDP is measuring–production not consumption.

Along the same lines the common comment that consumer spending is two-third of output or production is also completely wrong.

Consumer spending is two thirds of total consumption, but it is zero percent of production.

5 Ted September 11, 2010 at 9:26 am

Why is it that the immediate response to any and every economic question is “regulate it”? Honestly, the fact that there aren’t perfect substitutes domestically in an economy is probably because they’re incapable of producing those substitutes domestically. This is a weakness from the global perspective, where foreign economies will benefit from those that restrict businesses from producing products locally (an example here is drug trade). The decision to regulate here and ban imports is not only extreme, but “breaks” the model. The lack of imports in that case is not because the products can’t be produced locally, but because the imports can only be obtained via underground economies (see alcohol ban and drug trade again). Regulation is not a way to tilt a model in a desired direction, but instead a way to bend the model around the requirements of the regulation. Why do so few economists (Keynesians come to mind) realize this?

6 lagged_variable September 11, 2010 at 9:43 am

Bob and axiom (and spencer’s second “double counting” post) are exactly right. To say it another way, consider the alternative. If exports increase and imports decrease a country’s GDP, and $1 of exports for one country is $1 of imports for another, then think about the consequences for measured world output. We could double the output that is traded (all else equal), and world measured GDP would not change. If you believe that imports have a negative effect on GDP measurements, then logically you must believe that specialization and trade cannot increase world productivity. Which I hope we all agree is absurd.

Of course, FYI could still call this “mercantilist” since it’s only exports that increase the measured value of a given economy (whereas imports only increase utility via consumption or future output via investment). But as Angus points out, it’s just an accounting identity.

7 Josh September 11, 2010 at 10:04 am

Earlier this week there was a NYT article bemoaning how Chinese subsidies of solar panel manufacturing have resulted in Chinese dominance of the solar panel industry.

The free trader in me says, if they want to pay us to buy solar panels, great! More solar panels for us, and we’re richer at their expense.

The guy in me who’s worried about persistent high unemployment says, we have fewer jobs as a result.

I don’t believe the Chinese get a permanent advantage in solar panels out of this, nor (as Andy Groves complains) in lithium battery technology.

But I wish there were some rule of thumb concerning when our economic concern should shift from the benefit to the American consumer, to the unemployment rolls.

8 Jim B. September 11, 2010 at 11:24 am

Right. In an accounting sense, imports reduce GDP, CETERIS PARIBUS. However, the ceteris is seldom paribus. Countries that import more also tend to export more, and, on average, the two will be a wash on GDP, in the long run. Empirically (and theoretically), countries that TRADE more (imports and exports) have higher GDPs and higher GDP growth rates, on average.

9 Paul September 11, 2010 at 3:01 pm

Was it Krugman who pointed out, back in the eighties, that overly simplistic neo-liberal models of free-trade are bogus because of the importance of clustering in economic production and in the importance of the productivity of workers? The problem with free trade is that technology matters, but technology(and by that I mean methods of production and organization of factors of production) resides more or less in people, cultures and institutions. Mercantilism works, and it works because it results in a transference of technology and the establishment of new clusters of firms and the accompanying culture that supports phenomenal productivity. Refusing to play this mercantilist game is akin to agreeing with China that we will let our per capita productivity growth fall so that they can catch up with us.

10 Craig September 11, 2010 at 8:59 pm

“we just measure final consumption and this also counts the auto, steel, iron ore, etc., etc., just once.”

Yes, and this is what gives rise to the fallacy that consumer spending “drives” the economy.

A pox on Keynes’s house.

11 Bill September 11, 2010 at 10:37 pm

National Income Accounting likely overstates the value of imports by multinationals which have employed transfer pricing techniques to minimize US taxes, and understates the value of exports of intellectual property by MNCs from the United States. Until you get a grasp of how valuation is conducted–and how it is distorted–using the declared value of imports you will always overstate import transactions to a parent company. To give an example, let’s say I have a patent or a trademark; I value it and transfer the patent to a foreign sub in Ireland. If I undervalued the patent or trademark (and even if I fairly valued it) what happens is that the foreign sub licenses back the technology to the US sub under a license agreement and contracts with the US sub to perform R&D, getting the patentable results from the contract. The foreign sub takes license fees and uses them to finance patent improvements back in the US, and values the improvement patent it gets back from the US development at a much higher fee. The improvement patent is licensed back to the US at a very high license fee, reflecting the import value of an intangible asset. The US sub, paying the high license fee, shows little profit, and has lower taxes. Even though the invention is done in the US under an R&D development contract with the foreign sub, the value is captured abroad by the foreign sub, which later licenses back the invention to the US. The value of the export contract–the R&D development contract with the US sub–just covers the costs of development, and not the value. The Value of Imports to the US increases from the license of the foreign to the US sub, even though all the work has been done in the US.

Frankly, I question the idea of identifying the locus of value within an MNC, and thus even the value of National Income Accounting when it involves transactions and licenses between US and foreign subs. GIGO. And, to the extent we become more of an innovation country, the misplaced reliance on invoice, license or other transactions between related parties will only make National Income Accounting more distorted.

12 Six Ounces September 12, 2010 at 1:11 am

Sorry, I meant to say, “The value of all final goods and services…”

13 Ricardo September 12, 2010 at 4:30 am

GDP should be negative, right? The economy produced nothing, and it consumed goods and services for which it must either pay out of savings from previous GDP or borrow from future gdp.

Bob got this right before. GDP would be zero because all the goods and services consumed would show up under consumption. Then you subtract off the value of imports and are left with GDP of zero. This is the right answer because nothing was produced using factors of production physically located within your hypothetical country.

14 axiom September 12, 2010 at 6:24 am

“GDP should be negative, right?”

GDP is zero.

Income is zero.

Saving is negative, = (C).

Change in net worth is negative, = (C).

15 liberalarts September 12, 2010 at 1:59 pm

@MikeP-foreign investments in America do not directly cause a change in the trade balance. The purchase of American securities shows up as on the capital account. If the American that sold the investment to the foreigner uses the revenue from selling the investment to buy a Mercedes, then that adds to the trade balance deficit.

16 R. Richard Schweitzer September 12, 2010 at 8:22 pm

Did anyone think to go back to the basis for Colin Clark’s original statistical work on aggregate national production and the purpose for its organization?

17 Jonah Thomas September 12, 2010 at 10:03 pm

“Mercantilism works, and it works because it results in a transference of technology and the establishment of new clusters of firms and the accompanying culture that supports phenomenal productivity. Refusing to play this mercantilist game is akin to agreeing with China that we will let our per capita productivity growth fall so that they can catch up with us.”

Whether mercantilism “works” depends on your values, right? It’s kind of like a potlach. Mercantilism says you can get away with selling at low prices, giving stuff to your trade partners. From one point of view, why should anybody object? If the Chinese give us consumer goods at low prices, it’s like we’re getting stuff for free. We’re better off. When they give us a free lunch, why turn it down?

In another sense they feel like they win when they get the market share and they get the gold etc.

I remember a fragment of a dream. In the dream I’m competing for the Olympics triathlon, and I have a good shot at it. But there’s one competitor I’m pretty worried about. So I arrange that he gets three bavarian cream eclairs, or sometimes Napoleons, delivered to him every day. For free. So I go on training as hard as I can, and I meet my competitor and he has a little paunch. He tells me about it. “Free pastries! I’ve never had it so good! Mmm, those things are great, and completely free! Whatever else happens I’m coming out ahead on the deal.”

It’s all about your values, about what you think is important.

18 MikeP September 13, 2010 at 12:29 pm

Where I see this argument break down is this: There are around half a billion under-employed peasants in China. What products would China rather import from us, rather than allow those peasants to make for them?

Airplanes, software, product design, financing, etc., etc., etc — not in general the low end commodities you cite.

The US is at the high end of the value chain. China’s industrialization is bringing it into the middle of the value chain as they try to find the manufacturing volume to move people from farms to cities. Why do you imagine that the relationship between the US and the still-peasant segments of China isn’t completely complementary — exactly where comparative advantage offers the greatest leverage.

19 MikeP September 13, 2010 at 5:25 pm

I fail to see how your comment at all addresses your question, “What products would China rather import from us, rather than allow those peasants to make for them?” It appears to be a simple statement that trade deficits are prima facie bad — the very point being debated.

If you depend only on high-end products for your bread and butter, after awhile you may not have bread.

You trade for bread. That’s the whole point. And the poorer another economy is, the better off both of you are trading what you produce for the other economy’s bread.

So for example if Ford had concentrated too much on their high-end cars, they would not have had a timely hybrid.

“High end of the value chain” doesn’t mean “expensive”. It means “high value-add”. The hybrid is high-end.

20 MikeP September 13, 2010 at 11:50 pm

At this level it’s only a question of values. Do you care about employers and workers more, or do you care more about consumers? Whichever side you favor decides whether you think cheap imports are good or bad.

I don’t think those are the relevant values to compare.

At the economic level the values to compare are which of the two choices results in the wealthier economy. The answer here is clear: except in extreme corner cases that have to be maintained by impossibly intelligent and balanced government policy, the free trade option results in the wealthier economy.

And at the moral level one has to recognize that nations don’t trade: individuals and mutually voluntary collectives of individuals (e.g., corporations) trade. And if one individual or corporation has a trade deficit with another individual or corporation, then that is simply the business of those individuals or corporations and no one else.

The bottom line: To maximize its economy’s wealth and freedom, government must do whatever it can to maintain free trade without caring about deficit or surplus. This can be tricky if it insists on controlling the money, as most governments do today. But if government tries to control trade to advance some politically imagined ends and may be able to point to seen improvements due to such policy, the unseen damage eats away far more.

21 Jonah Thomas September 14, 2010 at 2:25 am


“At the economic level the values to compare are which of the two choices results in the wealthier economy. The answer here is clear: except in extreme corner cases that have to be maintained by impossibly intelligent and balanced government policy, the free trade option results in the wealthier economy.”

You have not described your reasoning, but you are quite close to a giant fallacy that I want to describe even though it might not fit your thinking.

Some people believe that free trade maximizes wealth. But there is no basis for this belief. Free trade may not maximize anything, or it might maximize something-or-other that you would not particularly want it to. As a result there is no guarantee that free trade will result in a wealthier economy. And yet I usually see no reason to predict that less-free trade will result in a wealthier economy. So for the sake of argument I will accept your hypothetical claim.

“And at the moral level one has to recognize that nations don’t trade: individuals and mutually voluntary collectives of individuals (e.g., corporations) trade. And if one individual or corporation has a trade deficit with another individual or corporation, then that is simply the business of those individuals or corporations and no one else.”

Here you insist that there are no consequences. But when a business goes into debt that can make a difference to its suppliers and its customers, to its other creditors, and in extreme cases to its debtors. And of course it makes a difference to its government. To say it is nobody else’s business is to say that there is no economy, only a collection of monads.

“The bottom line: To maximize its economy’s wealth and freedom, government must do whatever it can to maintain free trade without caring about deficit or surplus.”

I can see that under your assumptions that will maximise wealth. But I will look at some of the implications. One way this can work out is that when there is a trade deficit after awhile the money gets devalued. Then imports are more expensive while exports are cheaper, so the trade imbalance tends to be corrected with no action required by the government. But if for some reason the trade imbalance continues? The money must be devalued again. It becomes harder and harder for the poor or middle class to buy things in that country, because they cannot afford imports and they cannot afford their own products either — foreigners can pay in hard currency and they cannot. And if the rich still keep importing and the trade imbalance continues? Then increasingly the people figure out that they can live much better if they can reach other nations. An MD might make a better living as a motel maid elsewhere. The factories get bought and shipped elsewhere. The railroad rails get pulled up and sold as scrap. In the extreme case, no one wants to invest in this nation and everything of value crosses the border, because the best use of the national resources is to send them elsewhere.

Since this nation is not a good place to invest, the best thing its government can do is to systematically invest in more prosperous nations, stripping their own nation of everything of value.

Imagine that this was true for the USA. That our vaunted hi-tech industries were simply more expensive to operate here than in china, that the best use of every investment dollar was to invest it in china. Would it make sense for US government decision-makers to go through with that? What would we do with our own labor force? We couldn’t very well ship them all out to be guest workers in other countries.

Maximizing the total wealth is not the only goal.

22 MikeP September 14, 2010 at 2:51 am

But if another nation commits economic warfare against us, would we be better off by doing nothing or by making some sort of response?

Doing nothing, of course. Unilateral free trade is the wealth maximizing policy for an economy.

If I’m an author who writes books about the economics of free trade and you are a farmer who prefers protectionist policies, do you think I shouldn’t buy grain from you because you don’t buy my books?

23 Jonah Thomas September 14, 2010 at 2:49 pm

“First of all, what do you think foreigners do with all the dollars they earn building trade surpluses with the US? They invest them in US holdings — including holding dollars. A large part of the trade deficit is because the US is highly productive, highly stable and very inviting to investment. If instead, as you suggest, those holdings are material objects that cross the border, that’s trade surplus, no?”

And this is part of why we had the housing bubble? Our bankers found a way to package mortgages that those foreigners could buy, and they bought them in preference to the sort of investments that could make new useful products?

And our bankers invented lots of new “investment” vehicles that somehow crashed? Because we didn’t have enough actual honest investment opportunities….

“Your grim storyline simply doesn’t happen so long as no one does anything stupid, such as use state power to protect prior valuations of currency.”

What if a foreign power protects prior valuations of our currency. Many of the same bad things happen to us that would happen if it was our own government doing that “stupid” thing. If they want to hurt us, can they do it that way?

Imagine that China wanted to bomb our industry. Imagine the costs to them of a shooting war that accomplished that goal.

And now they get the same result just by protecting our currency, at far lower cost. Us factories closed down, deteriorating, not getting the upgrades they’d need to stay competitive.

It looks similar to me.

People say, don’t start a war if you can reasonably avoid it, because the enemy will fight back and it will probably cost more than you expect. And similarly they say, don’t make government economic policies designed to enrich your businesses at the expense of foreign businesses, because foreign governments will retaliate.

But when a foreign government does start a war with us, like Pearl Harbor, we go to the same inept government we don’t want to trust with anything else and we want them to fight back. We’ll pay whatever it costs to win the war. We don’t let foreign nations get away with attacking us.

And on the other hand, when they set up policies designed to wreck our economy, you say to ignore it.

You would not ignore it if they sent in saboteurs to bomb our factories. But when they use economic methods — some of the same methods our own government might use that have the effect of shutting down our factories and leaving our workers unemployed — you say to just let them.

24 Jonah Thomas September 15, 2010 at 6:22 pm

“I actually see no way that someone else independently supporting your own currency could possibly harm you. ”

I strongly suggest the following link:


25 Jonah Thomas September 16, 2010 at 3:34 am

MikeP, as usual you have completely missed the point.

You are making firm predictions about the behavior of complex systems based entirely on your intuition and that of people who have convinced you, without doing simulations to see how well your assumptions hold together.

You are certain for example that one national government can severely damage its own economy by manipulating its currency, for example, but that no other national government can damage the first nation’s economy by manipulating its currency.

You say that a central authority cannot predict a complex system. And I say that very likely you are also not very good at predicting a complex system. But you appear to be absolutely certain that you know how they work, when you have demonstrated no reason whatsoever for your confidence.

26 Jonah Thomas September 16, 2010 at 3:24 pm

“Yet only one of us is advocating a policy that requires predicting a complex system.”

No, neither of us is.

“Yes, I may have confidence in the natural laws that run economies and thus predict that free markets will produce better outcomes than controlled markets.”

Why not run simulations and get some idea under which circumstances free markets function well?

And then run simulations on the results of government interference in free trade. Is it true that governments always damage their own economies and nobody else’s? Is it even true that governments always damage their own economies more than anybody else’s?

See, if it’s true that china will inevitably damage their own economy more than ours, while we would inevitably damage our own economy more than theirs, then it follows that if they try to attack us we should ignore them completely. Maybe when they get tired of hurting themselves they’ll stop. But I have no idea how you got the idea this is true. It doesn’t seem at all plausible.

But it could be true, I can’t say I know for sure it isn’t true. Complex systems are hard.

27 Jonah Thomas September 16, 2010 at 10:50 pm

“Why not run simulations and get some idea under which circumstances free markets function well?

“And then run simulations on the results of government interference in free trade.”

Because the very notion that one can simulate complex systems of billions of individuals each pursuing their own goals — goals with complex feedbacks all their own — is laughable.

Well see, you make firm definitive claims about those systems, claims that you cannot even test. Now you imply that your claims are completely untestable, even in simulation.

This says that you know nothing about the topic you claimed to understand.

Look at the recent failure of macroeconomists to predict even the current unemployment rate. Look at their utter inability to even tell what the effects of government stimulus have been. The only thing they can tell you is that their models predict something. There is no attempt to root the models in reality or even test against reality.

That’s silly. Of course their models aren’t good enough to predict real numbers. For one thing, the numbers they start out with are not very good. For example, the underground economy is clearly large enough to matter. But our estimates about how big it is, depend on the very people who are supposed to persecute it. They are not very good estimates.

How much cash does the USA lose per month to international drug dealers? Who would you go to for an estimate? Various nations have legal restrictions on oil sales, and they generally find ways to cheat on those restrictions but the official figures do not include the smuggling. And the banking industry has still not been re-organized enough for anyone to understand it.

Weather models are now good enough that 5-day predictions are pretty good and 10-day predictions are often good. What kind of weather predictions would we get if much of the data was compiled quarterly, and the rest dribbled in at random times? And we were supposed to predict the weather 90 days ahead?

But you still have an opportunity here. You have a set of economic concepts based on first principles, that you use to get a variety of almost deductive proofs from. If you model your assumptions you could see whether they actually give the results you predict they should.

That might not tell you anything about reality, but still you could learn a lot about your beliefs. Your assumptions about economics themselves form a complex system and if you simulate them you might likely learn something about them.

28 Jonah Thomas September 20, 2010 at 6:30 pm

MikeP, you are admitting to rational bankruptcy.

When you state your assumptions, and you do simulations that show the consequences of your assumptions, you have more than when you state your assumptions and then provide “full on proofs”.

Depending on where you mark the start, we have well over 200 years of economists doing “proofs”. And the result is the economic thinking we have today. We have perhaps 50 years of OR guys doing simulations. Compare the progress of the two.

Somehow you prefer to study economics, a complex interconnected system, without simulations. Adam Smith had an excuse. He didn’t have the concept and he didn’t have computing tools to make it easy.

What’s your excuse? What possible excuse could there be?

29 MikeP September 21, 2010 at 2:01 am

To be fair, there probably is a place for simulation in some economic scenarios.

A couple decades ago I did do some simulation of market processes, and it yielded useful measured results that would have been difficult to compute. But those were automatons behaving as agents in a market to try to reach a best solution to a problem. The agents had simple budgets and a single dimensional understanding of values. They were merely using markets to optimize a solution.

Compare that to the dozens of dimensions of values and goals among millions of individuals with their wildly different preference orderings and even different marginal differences in their preference orderings. Simulation of this system may yield interesting metrics that might be useful for something. But it certainly won’t yield any fundamentals different from those assumed in the model. Thus you will not “learn a lot about your beliefs”: you will only reproduce your beliefs, adding false evidence and confidence where none is warranted.

Comments on this entry are closed.

Previous post:

Next post: