Recent Posts

Wednesday, January 18, 2017

Trade And Growth

Chart: U.S. Trade Balance

This article is a brief discussion of the relationship between growth and trade deficits in the United States. The previous two decades saw a marked deterioration in the U.S. trade balance, as depicted in the chart above. As a result, trade acted as a counter-cyclical brake on growth. One interesting property of the current cycle is that the trade deficit has been relatively stable when compared to that experience.

(This article is somewhat brief. I was ambushed by various software gremlins, and I have been getting ready to ship Abolish Money (From Economics)! to my copy editor.)

Imports Subtracting From Growth

One of the more awkward things to deal with in economics is the fact that imports subtract from gross domestic product (GDP) in the standard GDP definition. The equation is:

GDP = C + I + G + (Ex - Im),

where C = household consumption, I = business investment, G = government expenditures, Ex = exports, and Im equals imports.

Why imports have a negative sign in front of them is easier to understand if we discuss trade in services.

As an example, imagine that a consumer is buying a $100 piece of software that is downloaded from a software firm's web site.
  • If the firm is located in the same country, it is fairly straightforward that the purchase adds $100 to GDP. We can calculate GDP in a few different ways, and one way is to calculate it based on purchases. In this case, the transaction adds $100 to consumption expenditures.
  • If the firm is foreign, it is clear that nothing is being produced domestically (it would show up in the producer's home country GDP). If we calculate GDP by adding up expenditures, we need to adjust for this. In this case, the transaction will add $100 to consumption, but also add $100 to imports -- which enter with a negative sign into the GDP equation.
(The situation is more complex if we are discussing goods that are imported. Imported goods generally enter into retailers' inventories before they are sold to consumers, and so we would also need to track those accounting entries.)

One key point to note is that we could not have negative GDP as a result of imports; the subtraction is just cancelling out consumption of goods and services that came from external production.

Are Imports Negative For Growth?

Whether or not imports are bad for growth is a point of dispute. Noah Smith wrote a recent Bloomberg article in which he attempts to debunk the belief that imports subtract from growth.  Most of the article is running through the cancellation effect I described above. The more debatable point is whether a drop in imports would increase growth.
  • If the imported goods in question were a substitute for domestically-produced goods, GDP would presumably rise ("all else equal"). Consumers are budget-constrained, and so the substitution would not immediately change total consumption. (As time passes, greater domestic production would raise wage income, and boost consumption.)
  • If the goods were complementary to domestic production, Noah Smith argues that the import reduction (as a result of tariffs, for example) would impede domestic production. 
I am not hugely interested in debating free trade orthodoxy, and so the rest of this article discusses the cyclical effects of trade on GDP accounting, without worrying about behavioural issues.

The Trade Deficit and the Cycle

Chart: Annual Change in the U.S. Trade Balance

The chart above shows the annual change in the trade balance (based on the quarterly exports and imports of goods and services used in the GDP accounts) for the United States. This series is roughly equal to the amount that is mechanically added or subtracted from nominal GDP by the trade balance over a one year period.

(The calculations needed to determine the contributions towards real GDP are more complex, and I am not entirely convinced of their usefulness in this context. My key objection is that a price spike in imported energy prices represents a shock to domestic nominal incomes, and trying to ignore the price change seems silly.)

One thing to keep in mind when interpreting these charts is that I am scaling the amounts by nominal GDP -- which was continuously growing during this period. Therefore, if the trade deficit was stable at 2% of GDP (a straight line in the first figure) it would represent a steadily growing deficit in dollar terms, and so the annual change (second figure) would be negative amount.

During the 1990-2010 period, the change in the trade balance generally acted as follows:
  • it was mainly negative during an expansion; and
  • it reverted to positive level around recessions.
As a result, the trade balance acted in counter-cyclical fashion (subtracted from growth during an expansion, and added to growth during a recession). The post-2010 experience is different, in that the contributions to growth are smaller, and are on average near zero. However, this may not be particularly surprising -- growth during the post-2010 has been tepid, and so we would expect the drag from trade to be smaller.

Concluding Remarks

This article just introduces the basic relationships between trade and growth; I hope to do more detailed analysis at a later date.

(c) Brian Romanchuk 2017


  1. Nice description of the process, but it's important to remember that it is just accounting which means that it is translated on paper into a reporting currency according to artificial boundaries. Those are a definite as a result of the accounting policy - which determines the shape of the accounts. If you have different policies you get different accounts and a different shape.

    It's at the trade boundary between countries that the impact of accounting policy on a national basis start to take effect.

    Since most economists can barely draw up a balance sheet, it's hardly surprising that they can't get their head around the impact of a reporting currency and border definitions in accounting policy.

    And that's before we get into the weaknesses of the GDP measure.

    So we've always got to be careful and remember that the accounting policies used aren't the only ones that are available. In fact they might be wrong as well.

  2. JW Mason has a nice post showing most trade differences come from demand.

    I've been thinking that what's missing from Current account surplus / deficit discussions is Minsky. A trade surplus will go straight to firms as profit / savings. This will both support animal spirits and a greater level of leverage driven investment.

  3. I think that this article - and Smith's - misses the point about BoP constraints.

    BoP constraints are not thought to kick in in the short-term. Obviously 'M' is not sucking income out of the economy like some sort of black hole.

    Rather they are thought to play out in the medium-to-long term. In concrete terms, over the course of maybe 5-10 years.

    The way that this works in a floating FX system is through devaluation. Eventually a country running a substantial trade deficit will face depreciation**. When this depreciation occurs the price level will rise as imports are rendered more expensive. This results is inflation and an erosion of real living standards. At this point anti-inflation policy tends to kick in which hampers economic growth.

    In theory you could not run anti-inflation policy and just let the inflation climb. But this would either lead to a wage-price spiral or (more likely) an erosion in real wages. The former will require action. The latter will result in lower living standards.

    Of course, the floating FX fans will say that the BoP will simply rebalance. Sure, M will fall; but X will rise. So, they will say, the country will gain with better, higher paid jobs as manufacturing employment rises.

    But this is a leap of faith for which there is little evidence. It depends on the Marshall-Lerner conditions holding and they do not hold in most countries with a large tradeables sector. As someone quipped in the House of Lords the other day "If currency depreciation leads to a more competitive economy then the UK since 1945 should have become the most competitive economy in the world!" Indeed.

    ** Perhaps not if you are a global reserve currency. And that may mean that the US is immune from all of this... WHILE IT REMAINS A GLOBAL RESERVE CURRENCY.

    1. I am not attempting to write about fancy concepts like balance of payments constraints here; I am just writing about the straightforward mechanical GDP accounting growth effect. I only linked to Smith's article to reference the small literature that says these mechanical effects allegedly do not exist (because trade is good!). For example, I found an article that said "imports subtracting from GDP" is a Keynesian myth...

    2. Fair enough. But I don't think that you can talk about trade and growth without these more complex discussions.

      For the most part I don't think they seriously impact the US due to the dollar though. For now.

    3. It acts as an automatic stabiliser, which was a point that I was going to take up in a later post. It's analgous to things like unemployment insurance - we can discuss how they stabilise growth rates in the short term without worrying about the long-term ("steady state") growth effects.

    4. I'm certainly not sure that I am referring to long-run (structural) impacts. I'm talking about a potentially unstable element that causes crises -- not unlike, say, Minskyian debt bubbles.

    5. Crises would be related to the financing of trade flows; the trade of real goods and services are unlikely to cause the crisis. (I am lumping a potential currency collapse under "financing.") I mainly discuss developed economies, and I am hard-pressed to think of any trade-related crises in floating currency sovereigns.

      Like other people in the MMT camp, I definitely do not spend time worrying about the external constraint that other post-Keynesians emphasise (which is why I typically describe myself as being in the MMT camp versus "post-Keynesian").

    6. There's a ton of such crises in developing markets. There have been one or two in the past two years.

      In the developed markets they tend to be more slow-burn and less immediately obvious. The most recent was in 2007-08 in the UK when the sterling collapsed 25% or so. I believe that we are about to see another this year in the face of the Brexit.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.