Do worsening trade deficits really slow economic growth? I’ve been making this point for quite some time in the case of the United States, but a recent post on America’s resurgent trade shortfall undermining the current recovery sparked some major push-back from a friend of mine who’s a senior official at a major global institution. Since he made his arguments to me privately, I won’t disclose his identity. But that doesn’t prevent me from dealing with his points in detail.
My post was based on the methodology used by the federal government since the 1930s to calculate not only overall U.S. economic growth, but the sources of that growth. The data series I used – the gross domestic product (GDP) – isn’t the government’s only way of measuring and analyzing growth. But it’s certainly the most widely used and cited, and since the end of World War II, it’s become an international standard.
As explained by the U.S. Commerce Department, GDP shows us “the total value of final goods and services…produced by the U.S. economy” and it does so by measuring “final purchases by households, business, and government by summing consumption, investment, government spending, and net exports.”
This phrasing indicates a peculiarity of the GDP – it gauges production indirectly, by adding spending totals by major segments of the economy, and then subtracting imports, since purchasing them amounts to purchasing goods and services made outside the United States. Nonetheless, when spending in one of those GDP categories increases, the effect on growth is to increase it. When spending in one of those categories shrinks, the effect on growth is to reduce it.
The logic behind that methodology also convinced the GDP’s originators, and those who have used it since, to score a worsening of the trade balance (whether an increase in a deficit or a decrease in a surplus) as a growth-slowing development, and an improvement (whether a decrease in a deficit or an increase in a surplus) as a growth-enhancing development.
The indirect approach taken by the GDP seems to be one reason that my friend contended that, “The relationship between trade flows and GDP in your blog is an accounting one, not the effect of trade on growth.” Yet it seems that the conventional wisdom of the economics profession disagrees with him. According to no less than the late Nobel prize-winner Paul Samuelson, who wrote the history’s most influential economics textbook, the GDP enables Americans (and others of course) to judge whether the economy is contracting or expanding.” Doesn’t the same hold for GDP’s building blocks?
Not that they’re infallible, but government economists plainly believe that it does. The Commerce Department still routinely issues statements such as “The deceleration in real GDP growth in the fourth quarter primarily reflected an upturn in imports, a downturn in federal government spending, and decelerations in nonresidential fixed investment and in exports that were partly offset by an acceleration in PCE, an upturn in private inventory investment, and an acceleration in state and local government spending.”
Indeed, in every one of its quarterly reports on the gross domestic product (including the revisions), Commerce publishes a table (number 2) specifying how each of the components of GDP, including trade, has contributed to or subtracted from whatever change the total GDP has undergone in recent quarters and recent years.
Moreover, it seems equally plain that these government economists believe that these effects appear in the real world, not simply on abstract or symbolic balance sheets. Here’s Federal Reserve Board economist Steven B. Kamin discussing obstacles to America’s economic recovery at a Federal Open Market Committee meeting in March, 2009: “U.S. exports plunged in the fourth quarter and were a significant drag on U.S. Growth.” At that same meeting, then Fed governor and now Chair Janet Yellen voiced similar concerns. And I could cite hundreds of similar examples.
Nonetheless, because even the most famous economists make mistakes – often whoppers – my friend’s accounting point is worth examining on the merits, too. When I responded to his original criticism by pointing out that all the components of GDP are vulnerable to the charge of measuring output only indirectly, not just the trade components, he agreed that “That’s the point. [GDP] only refers to spending, not growth and employment, which is what we care about.”
Judging from some other critiques of GDP (and remember, his comments were constricted by the tight character limits of Twitter), he might be referring to the fact that technically (and importantly) speaking, a change in the trade balance is anything but an automatic indicator that a company in the United States has shut down some of its production. Even if sales in the United States are lost because domestic customers substituted imports for domestically produced goods or services, or export sales are lost because foreign rivals similarly won out, production of that domestic good or service doesn’t immediately, or necessarily ever, fall.
For example, the production lines could easily be kept open in hopes of finding new customers for that output, either at home or abroad. In addition, many and indeed most American producers presumably would doubtless still retain domestic customers, and many of these producers could sell more to these domestic customers, or find new ones. At the same time, those sales would have nothing to do with trade, and would be included instead the domestic components of GDP.
To complicate matters further, a decrease in imports or an increase in exports need not impact production, either. For example, the new customer could be supplied from existing inventories (although inventory changes are tracked in a separate GDP account). Alternatively, although imports may be falling, domestic conditions may be deteriorating, too, meaning that production levels need to stay roughly the same because overall demand for this or that product isn’t changing much. Yet as the critics have noted, although the effects on output are so uncertain, and may not emerge at all, the calculation of GDP’s increase or decrease does need to subtract worsening trade deficits from the final figure in order to avoid make sure that only domestic output is counted.
So are the critics right? Well, that’s a legitimate conclusion. But here’s the rub. As I suggested in my response to my friend, many of the same objections could be made about all the spending-based categories of GDP and how they affect growth. For reasons like those cited above, increases or decreases in consumer spending or business spending or government spending have no inherent effect on production levels of anything, especially over relatively short periods of time. So if we’re going to stop talking of improving or worsening trade balances affecting growth rates, we need to stop talking about rising or falling personal or business or government spending levels affecting growth as well. But if we do, then it’s hard to understand how Americans or anyone else could understand with any precision what their economies’ growth engines and growth obstacles are.
Moreover, even if changes in trade flows don’t automatically change output levels, especially in the near term, they’re likely to have effects over the longer term. It’s certainly true in theory, for example, that industries getting swamped by imports at home could find enough new customers abroad to offset the losses, or more. Further, despite surges of imports, their home markets could be growing so strongly that enough business emerges to make everyone happy. But how plausible do scenarios like those sound to you? And again, changes in domestic markets tell us nothing about the contributions trade flows per se are making to economy’s growth or about their contractionary effects.
My friend, however, seems to think he has answer to these ripostes, too. He has written that “Producers produce gross output; GDP measures value added only. Employment depends on gross output. Import[s] can increase output.” But on the first point, he seems flat wrong,and on the second, he’s on shaky ground at very best.
As per the statements above from the Commerce Department, GDP tries to measure the final value of all goods and services produced in the United States. Value added – a different gauge of output – is measured by a different set of statistics. They seek to avoid the double-counting of production that would be created by counting the value of the parts and components and other inputs that make up most goods (especially in manufacturing), along with the full value of the final good. So for example, in the value added data (and this description simplifies to some extent), the figure assigned to an automobile would be the difference between the sum of the value of its parts and the value of the final vehicle.
Value added, therefore, is in many ways a superior methodology than gross domestic product. Unfortunately, these data don’t contain any figures on trade flows. So they don’t enable any calculation of trade’s effect on growth.
Regarding imports creating more output than would otherwise exist, that’s possible in principle, but implausible on net for the U.S. economy specifically. The idea is that many imports are precisely those inputs of final products mentioned above. If they are cheaper, or of higher quality, or both, compared with their domestic counterparts, they could in theory make the final products so much more competitive that their use would greatly juice sales.
What this argument overlooks, however, is that a huge share of U.S. manufacturing production in particular that consists of these various inputs – which range from chemicals to ball bearings to semiconductors to fasteners and dozens of other products. For the imports-create-output argument to work, then, replacing domestic inputs with foreign would have to generate an amount of final products that’s greater than the amount of U.S.-made parts and components and materials that are displaced. Again, that’s possible. But the numbers tell us it’s highly unlikely.
Here’s a good compromise solution. I’ll agree to stop saying that growing trade deficits reduce the rate of economic growth (and vice versa), at least over the short term, if my friend (and others) stop insisting that changes in trade balances have no impact on growth at all. A better way to look at the matter would be to note that trade balances and their changes measure the efficiency (in terms of generating domestic output) of America’s spending. The more the nation imports with a given level of spending (whether by consumers, businesses, or government), the less spending gets recirculated back into the domestic economy to generate more demand for American-made goods and services.
Given that more than five years after the last recession ended, the United States is still starved for growth, I don’t see how the conclusion can be reasonably avoided that the trade deficit’s simultaneous comeback points to a serious problem either with Americans’ spending patterns, with their trade policies, or with both.