U.S. Trade Deficit

Since the end of 2012, the yen has depreciated significantly and the dollar has appreciated significantly. Enough time has passed to look at how U.S. and Japanese exports have responded to these exchange rate moves. Both countries' exports are also shaped by global developments—the oil/commodity price shock of 2014 slowed emerging markets' import demand, for example. But there should still be information content in the relative performance of U.S. and Japanese exports. And guess what, exchange rates matter. Certainly the different trajectories of the dollar and yen would seem to provide the most straight-forward explanation for the relative performance of U.S. and Japanese exports (using the cumulative contribution of exports to real GDP growth as the measure of performance).* The same point holds for net exports—the overall trajectory of Japanese and U.S. imports has been remarkably similar, though I suspect for somewhat different reasons (Japanese import growth has been weak after 2014 because of weak consumption growth, U.S. import growth has been weak because of weak investment growth—and an inventory correction, though there are now signs U.S. import growth is picking up). Net exports clearly have helped Japan and hurt the U.S. over the past several years. If you step back a bit and plot the cumulative contribution of net exports to Japanese and U.S. growth since say the end of 2004 (there is no perfect starting point) there is a decent (inverse) correlation between cumulative growth in net exports and changes in the real effective exchange rate for both countries.   Note that for the U.S. the depreciation in 2003 likely had an impact on the size of change in real net exports after 2004 as well, given the lags. Net exports are once again up substantially since the end of 2004 for Japan (even with the need to substitute imported gas for nuclear power). And net exports are no longer up much for the U.S. since the end of 2004 (even with the dramatic reduction in U.S. net oil imports that resulted from the huge growth in U.S. production of light tight oil) The exchange rate swings of the last five years have changed the picture dramatically relative to say 2012.   The positive impact of the dollar's 06 to 13 weakness is fading from the data.   And the contribution from real net exports to Japanese growth in the past four or so years looks increasingly similar to the contribution of real net exports to Japanese growth back in 05,06 and 07—another period of yen weakness. The strong contribution to growth from net exports recently is one reason why Japan’s current account surplus is now running at close to 4 percent of its GDP ($184 billion in 2016).  The rise in Japan’s external surplus offset some of the 2016 fall in China’s current account surplus, keeping Asia's overall surplus up. And I think there is growing evidence that the depreciation of the yuan in 2015 and 2016 is also starting to have an impact on global trade. In 2015 and 2016 I suspect China's export performance was slowed by the lagged impact of the yuan's appreciation in 2014 (the yuan followed the dollar up against a basket). But now the impact of the 2014 appreciation is fading, and the 2016 yuan depreciation is starting to show. China's May’s export volume growth looks to be around 8 percent—in line with the average for the first four months of the year (China only reports y/y changes in volumes, and the numbers are distorted by the new year, so there is no perfect method of getting the number). That is likely a bit faster than the roughly 4 percent y/y growth in global import demand in the IMF's forecast.     My guess is that China’s export volume growth this year will exceed global export volume growth. As one would expect given the yuan’s depreciation since mid-2015. Exchange moves still tend to have predictable consequences. Chinese import volume growth continues to be fairly strong (April was relatively, but May looks solid). But as China’s credit tightening starts to bite, I would not be surprised if China’s surplus also starts to rise— * The q1 2013 jump in Japan's exports was matched by a parallel jump in Japan's imports.  The symmetrical revisions seem to reflect from a revision in the national income product accounts.  I did not revise the numbers to strip out the effect.  It has no impact on the trajectory of net exports.  

China tends to import natural resources—oil, iron ore, soybeans and the like. Now that China has a large industrial economy, it doesn't have much choice: China wasn’t blessed with a ton of oil or gas, or a ton of arable land (relative to its population). And China tends to export a lot of manufactures relative to its own imports. It isn't just that the cargo ships that sail across the Pacific often struggle to find cargoes for the return trip—the trains that have been set up to send China's manufactures to Europe as part of China's highly touted One Belt and One Road Initiative also often return empty. Jörg Wuttke wrote in the Financial Times last week: "The planned rail link between central China and Europe, which is undergoing trials, highlights the challenges: five trains full of cargo leave Chongqing for Germany every week, but only one full train returns." Wuttke's anecdote captures something important about the nature of manufactured trade with China. If you net out imported components for reexport, China really doesn't import many manufactures: manufactured exports are about 12 percent of its GDP, and imports—net of processing imports—are just over 4 percent of GDP, leaving China with a large surplus in manufacturing trade. Exports as a share of GDP have come down after the crisis, but so too have China's imports of manufactures for its own use. Some imbalance here is normal: China naturally will trade its manufactures for the world's commodities, and run a manufacturing surplus. But China's the size of China's manufacturing surplus has created something new, given the size of China's economy. China's manufacturing surplus topped 10 percent of China's GDP at its peak before the crisis—so it has come down a bit. But its manufacturing surplus is still large. Since 2009, it has ranged from about 7.5 percent of China's GDP to around 9 percent of its GDP (thanks to China's recent stimulus it is at the low end of that range). Most of the fall in China's reported current account surplus after the crisis thus has not come from a fall in its surplus in manufacturing. Rather, much of the change in the overall balance reflects a swing in the income balance (better measurement of reinvested earnings, and a fall in interest income on reserves), from a rising commodity import bill and, recently, from rising tourism imports (excluding tourism, China isn’t a big importer of services). Bilateral number can mislead—so if you want to avoid the next bit, feel free. U.S. manufactured imports from China are roughly four times as big as U.S. manufactured exports to China (counting exports to Hong Kong, and using the NAICS definition of manufacturing, which includes petrochemicals). Adjust roughly for value added, and the discrepancy falls to maybe 3 to 1—as China’s exports have a bit more imported content than U.S. exports (OECD numbers here). In this case, I think the bilateral numbers illustrate something that is also globally true: China exports about three times as many manufactures as it imports, once imported components are netted out—its pattern of trade with the U.S. is actually fairly typical. And the U.S. also imports far more manufactures than it exports. Changing this basic pattern will be hard. On the U.S. side, tax cuts that push up the budget deficit won't help, especially if budget deficits raise interest rates and push the dollar up. And for all of China's new pro-trade rhetoric, China remains committed to developing an indigenous aircraft industry, an indigenous semiconductor industry, and a stronger medical equipment industry. In a host of sectors where China is now an importer, it wants to become an exporter: see Keith Bradsher's reporting in the New York Times. The new, highly touted "early harvest" deal with China doesn't try to change this pattern. It focuses on commodity exports (beef, natural gas) and increasing the ability of U.S. financial services firms to compete in China—not on opening new opportunities to export manufactures to China. That in part is why Dan DiMicco of Nucor is not a fan of the deal: “This is disappointing on many levels. We are rewarding China before stopping their massive trade cheating.” China’s willingness to trade beef for poultry isn’t new, or a surprise—in fact China had already agreed in principle to resume beef imports last year. In some ways, the most interesting bit of the agreement is the hint of a future agreement to increase U.S. LNG exports to China. The details aren't all that clear. The announced deal didn't actually do more than encourage Chinese energy companies to buy U.S. LNG: "The United States welcomes China, as well as any of our trading partners, to receive imports of LNG from the United States. .... Companies from China may proceed at any time to negotiate all types of contractual arrangement with U.S. LNG exporters, including long-term contracts, subject to the commercial considerations of the parties." As China already buys some LNG from U.S. liquefaction facilities in the spot market, the real impact of this will only come if China commits to a long-term contract to take up a steady stream of LNG. There are several additional points to make here. First, China doesn’t actually need to buy LNG to meets its domestic energy needs. China has no shortage of domestic coal, and its domestic coal is cheaper than imported gas. But gas is cleaner—and less carbon intensive. Now that global traded gas prices are down a bit, the cost is probably a bit easier to swallow. Second, Asia actually doesn’t need much U.S. gas right now. The Asian market is currently well supplied—thanks to large amounts of Australian LNG capacity that are now coming on line. Gas in Asia has traditionally been expensive because Qatar used its market power (Qatar has a ton of low-cost gas—it just didn’t want to flood the market) to keep prices high. But with lower oil prices and new competition from East Africa and Australia, Asian prices have come down. This is a problem for the U.S. LNG export story—the liquefaction is actually quite costly, and it isn’t clear that the current price differential is high enough to support a major expansion of U.S. gas exports. Clifford Krauss of the New York Times recently reported that some investment groups have all the permits they need to build more LNG capacity, but haven’t chosen to go forward because they haven’t found a buyer for the gas: I am not sure that China is committed to buy at any price. Nor is it clear that the U.S. is committed to allow Chinese investors to take an equity stake in a LNG liquefaction facility—a part of China’s preferred deal structure (see, for example, China's equity stake in the Yamal LNG facility in Siberia). Third, the implications of higher U.S. gas exports on the broader U.S. economy are ambiguous. The construction of a terminal no doubt generates demand for steel and for a range of high-end components, and raises demand for domestically produced gas. But exports also tend to pull up the domestic price of gas, and most Americans are consumers of gas (and of energy more generally) not producers. Exports thus make the average household worse off—even as they generate concentrated gains for the owners of gas fields, workers in the natural gas sector, and those workers who supply components to the liquefaction facilities. And to a degree exports of gas also trade off with exports of energy-intensive manufactures. The scale of the impact of higher exports on domestic prices is likely modest in the U.S. (it has been significant in Australia). U.S. gas supply now looks to be relatively elastic, so it won’t take much of a rise in price to induce the additional drilling needed to meet higher export demand. Especially as it still isn’t clear to me that China is willing to commit to long-term supply contracts at a price that would make investment in additional LNG export capacity economically attractive on a scale that would make a material impact on the market. But I do wonder—like Rory McFarquhar—whether the U.S. could have used China’s desire for a special relationship in gas as a source of leverage to try to get a bit more out of the negotiations in other areas, areas where China is less ready to give than in beef. LNG is as much a Chinese ask as a U.S. ask, and, well, the U.S. had a set of realistic non-Chinese alternatives. Using the gas at home as part of a domestic manufacturing strategy, for example. Or expanding pipeline exports to Mexico as part of a stronger energy partnership in NAFTA, or prioritizing LNG sales to allies in Europe and Asia, like Japan ... Bonus charts: Chinese manufacturing trade surplus (goods trade net of primary products), as a share of China's GDP: And the U.S. manufacturing and non-manufacturing trade balance with China and Hong Kong, as a share of U.S. GDP: Both sides of the imbalance fell a bit in 2016, as U.S. consumer goods imports were weak and China's stimulus pulled in imports. I though have questions about the sustainability though of both legs of the adjustment -- as U.S. imports seem likely to pick up after an inventory correction and China's weaker real exchange rate should support China's exports going forward while China is pulling back on its stimulus.

The U.S. trade deficit did not change much in the first quarter. The real (price adjusted) goods deficit that is. Oil prices were a bit higher in q1 than q4 (they are now back below their q4 levels, but that will impact the June and July data) The quarterly real non-petrol deficit has basically been constant—with a soybean specific story explaining the fall in the Q3 deficit. The non-petrol deficit excluding agriculture hasn’t moved much. Frankly I am a little surprised. I would have expected the lagged impact of the 2014 rise in the dollar (the broad real dollar is up 18 percent since mid 2014) to still be having a bit of an impact on the 2017 data—and also to see an ongoing drag on exports so long as the dollar stays at its current (appreciated) level. But that didn't obviously happen in the first quarter. Though I would argue that the real trade deficit is slowing getting bigger if you look past the impact of the q3 soybean export surge, albeit at a much slower pace than in 2015. The relative stability of the trade deficit in early 2016 came from weakness in U.S. real imports. That offset the weakness in US real exports. In the last couple of quarters, the story has changed a bit. Real imports have started to flow, but so too have real exports. A plot of the cumulative contribution of net exports to U.S. GDP growth since the start of 2014 makes this clear. Net exports over this time have, mechanically, subtracted about 1.5 percent from U.S. growth—with most of the deterioration coming in late 2014 and early 2015, just after the dollar strengthened. There hasn't been much further deterioration. Net exports knocked about 20 basis points off growth over the course of 2016, and were basically flat. I am inclined to stick to my original forecast, one based on the expectation that there is still a bit more deterioration in the U.S. trade balance baked in because of the dollar's past appreciation. I would expect a 18 percent move in the real exchange rate to raise the real trade deficit by a bit more than 2 percent of GDP over three or so years (a standard rule of thumb is that a 10 percent move in the real effective exchange rate changes the non-oil trade balance by between 1 and 1.5 pp of GDP; the Fed's trade model for example suggests that a 10% exchange rate move knocks 1.5 percent off net exports. That is also the IMF's central estimate).* I expect the trade deficit to widen a bit in the next few quarters. But for now the data shows a somewhat smaller expansion of the trade deficit than I would expect given the size of the dollar's appreciation. P.S. last year I thought the US real petrol deficit was heading up on a sustained basis. That forecast hinged on an assumption that $45-55 oil would keep U.S. production at best constant. The oil sector has outperformed—with a sizeable increase in drilling coming at a lower price point than I expected. I no longer expect the petrol deficit to rise. * Bill Cline of the Peterson Institute also projected a deterioration in the non-oil balance of around 2 percent of GDP. See his 2016 policy brief, and in particular figure A3.