Monetary Policy

There is no single definition of manipulation, to be sure—so no way of definitively answering the question. Over the last ten or so years, manipulation has been equated with "buying foreign exchange in the market to block appreciation." That definition is certainly built into the criteria laid out in the 2015 Trade Enforcement Act. But “buying reserves to block appreciation” wasn’t hardwired into the 1988 act, which has a much more elastic definition of manipulation. Yet even if the 2015 Trade Enforcement Act isn’t the only possible definition of manipulation, it still provides a bit of guidance - as President Trump implicitly recognized today: "Mr. Trump said the reason he has changed his mind on one of his signature campaign promises is that China hasn’t been manipulating its currency for months." The thresholds of being called out for "enhanced analysis" that the Treasury was required to set out in the 2015 act aren’t perfect—no measures are. The threshold for the bilateral trade balance is genuinely problematic. It lets small countries with a propensity to intervene in the foreign exchange market off the hook for one. And even if you think there is sometimes valuable information in the bilateral trade data (many don’t), the bilateral balance really should be assessed on a value-added basis.* But the current 3 percent of GDP current account surplus and 2 percent of GDP in intervention thresholds are certainly reasonable. Those criteria show that China should have been singled out for “enhanced engagement” from 2005 to roughly 2012, but not since. But all criteria can be gamed. And I worry a bit that China has been revising its current account data with the goal of keeping the headline external surplus down—it is hard to overstate the number of times the details of China’s services data have been revised since 2014.*** So Cole Frank and I looked at what would happen if some of the Treasury criteria were changed, or more realistically, augmented.**** For example, looking at China's overall goods surplus alongside the current account surplus might sense if you have doubts about the current account numbers. China's good surplus was a bit over 4% of its GDP in the last four quarters of data (4% of China's GDP is close to $500 billion, a significant sum). Or you could China’s surplus relative to global GDP to try to get a sense of how China's surplus is impacting its trading partners (a better measure is world GDP ex China, but doing that takes a bit more effort, especially if you intend to track more than one country). That also could put China’s current account surplus back in the zone of concern: But, as the Trump Administration now seems to recognize, there is no realistic way of getting around two facts: (1) China recently has been selling not buying reserves, so its intervention mechanically has served to limit the pace of depreciation, and (2) both China’s goods surplus and its current account surplus are heading down, so the Chinese overall policy stance isn't obviously impeding balance-of-payments adjustment. All this said, I would note that nothing tends to burn through reserves quite like a controlled depreciation. A significant portion of China’s reserve sales are a result of the fact that China has managed its currency in ways that helped generate the expectation that it wanted a gradual weakening of its currency. If China can stabilize expectations (and limit outflows from the state banks), its reserve sales might fall off fairly quickly.***** * There are data limitations here. The last OECD numbers on Chinese value-added in Chinese exports (around 70 percent) are from 2011. Things have changed since then, with Chinese value-added rising relative to China’s total exports. More and more high-end components are now made in China, even if they are not made in China by Chinese firms (see Bob Davis and Jon Hilsenrath of the WSJ; "China, once an assembly platform that sucked in commodities and manufactured goods from abroad, put them together and reshipped them, is now producing much of what it needs domestically. Benjamin Dolgin-Gardner, founder of Hatch International Ltd., an electronics manufacturer in Shenzhen, China, says he now uses Chinese-made LCD screens rather than ones made elsewhere in Asia for the tablets he produces. Memory chips for MP3 players are also made in China rather than imported from Japan and South Korea"). The famous Apple example now deceives—the relevant concept here is China’s share of total manufacturing value-added, not China’s share of the profit on the sale of an iPhone. Fun factoid: the New York Times reporting suggests that iPhones for sale in China are technically made offshore, in a bonded zone, then exported to Ireland (in a legal sense) before being legally imported into China from the bonded zone after downloading the needed software from Ireland (or wherever Apple now has parked its “global” IP rights)…an interesting example of modern global value chains. ** Japan and the NIEs (South Korea, Taiwan, Hong Kong, and Singapore) all also have larger overall external surpluses, relative to the size of their economies, than China now does. So in this case the bilateral value-added deficit maps to the imbalances in the global BoP. *** China's service deficit initially was in other business services. Then the 2015 tourism numbers were revised up, with jumps in tourism imports, exports and the tourism deficit. Then the revisions were extended back to 2014. And most recently the numbers were revised again, so not there is no upward jump in tourism exports -- only a jump in imports. The revisions have been linked to a broader change in how China collects its tourism data, with more reliance on electronic payments numbers and less on the numbers reported by China's tourism trading partners. There are of course innocent explanations for these revisions: almost all services data is a bit suspect in real time and is revised as better measures come in. But it is also clear that the net result of after all the various revisions was a big jump in the services deficit in 2014, just when the goods surplus was starting to rise on the back of Europe’s recovery and lower commodity prices. And given the methodology now used to calculate tourism imports has created a significant gap between China’s tourism imports and its counterparties' tourism exports, it seems likely that China’s current account balance now includes at least some financial transactions that really should be counted as financial outflows (such a reallocation would reduce the services deficit, and increase the current account surplus). **** The Trade Enforcement Act statute allows the Treasury to set out its own quantitative thresholds, but requires that the Treasury perform "an enhanced analysis of macroeconomic and exchange rate policies for each country that is a major trading partner of the United States that has-- (i) a significant bilateral trade surplus with the United States; (ii) a material current account surplus; and (iii) engaged in persistent one-sided intervention in the foreign exchange market." ***** I also have great confidence in the ability of China's authorities to engineer official outflows that would substitute for reserve growth should China start buying foreign exchange in the market. I consequently doubt that China will ever trigger the 2% of GDP in reserve growth threshold. But that is a topic for another time.

I confess that I probably am the only person in the world who—setting aside the internal politics of the Trump White House—would be excited to write the Treasury’s foreign currency report this quarter. Not because of China. I would say China met the existing 2015 manipulation criteria in the past and I would put the criteria under review (I personally think the bilateral surplus analysis should be complemented with value-added measures, which would reallocate some of China’s surplus to Japan, Korea, Taiwan, and the like). I might even find a way to warn that a country that guides its currency down can be guilty of manipulation under the 1988 law even if it is selling some of its foreign currency reserves (a controlled depreciation is hard, and usually requires reserve sales to manage the pace of depreciation). But not name China now. The U.S. would be completely isolated in naming China now, the impact of China’s 2016 stimulus seems to have been bigger than the impact of the renminbi’s depreciation and there is plenty of scope to get tough on China on other trade issues. As the Financial Times notes" "It is in no one’s interest, including the US, if Beijing suddenly stops intervening to defend the renminbi and a destabilising rush of capital flight and sharp devaluation follows." Rather, I would be excited to find ways of keeping the heat on Korea and Taiwan up, even if neither likely meets all three of the criteria in the 2015 Trade Enforcement Act. And, well, geopolitics probably is a constraint on getting too tough on Korea right now. It is often argued that countries won’t change their currency policies with a gun pointed to their heads, so explicit threats won’t work. Fair enough: threats do not always work (see the Freedom Caucus, health care). On the other hand, sometimes countries get a bit locked into a certain set of export promoting policies, and won’t change unless their feet are held to the proverbial fire. Korea for example still seems pretty comfortable intervening to keep the won in a band. It seems to have intervened again to limit the won's appreciation in late March (at around won 1115, the market has subsequently turned). And Korea's band is set in a way that keeps the won much weaker than it was before the global crisis, allowing foreign demand (via the trade surplus) to help offset the domestic impact of Korea’s weak social safety net, forced pension savings, and tight fiscal policy. And Taiwan has been comfortable with a weak New Taiwan dollar maintained in part through intervention and in part through encouraging Taiwanese financial institutions (life insurers) to invest ever larger sums abroad. The last Treasury foreign exchange report noted that Taiwan met the current account surplus and reserve accumulation criteria of the 2015 trade law. It escaped being dinged because of its small bilateral surplus with the United States. That report may have gotten Taiwan’s attention (I certainly tried to help the process along too). Taiwan’s intervention fell off a bit in the fourth quarter. And Taiwan is reported to have largely kept out of the market in the first quarter of 2017. That is one of the reasons why the New Taiwan dollar appreciated by about 5 percent against the U.S. dollar (and against the yuan—which probably matters more for Taiwan). As a result, Taiwan’s reported reserve growth (net of interest income) over the last four quarters of data (calendar 2016) dipped to right 2 percent of GDP (The Treasury threshold). That alone shouldn’t let Taiwan off the hook though. We don’t know how much Taiwan really intervened, because Taiwan doesn’t release data on its forward book (Taiwan has not voluntary adopted the SDDS disclosure standard, which it could do even if it isn't a member of the IMF). Taiwan claims it doesn’t intervene in the forward market, but, well, the old notion of “trust but verify” would seem to be relevant here. If Taiwan doesn’t have a forward book, it shouldn’t have a problem releasing the relevant data—including the historical data. Now the Treasury calculates reserve growth without explicit reference to the balance of payments data. It adjusts headline reserves for valuation, and it nets out estimated interest income. The Treasury's interest income adjustment—which is debatable, I personally think interest income should be sold for domestic currency, with the proceeds paying the coupon on domestic sterilization instruments and with any excess going into a reserve fund or sent back to the government—should help Taiwan. It should pull Taiwan's intervention below 2% of GDP in the Treasury's eye (the details of how you do the calculation matter it turns out). My gut is thus that Taiwan no longer meets two of the three existing Treasury criteria. So if Treasury sticks to those criteria, Taiwan likely has gotten itself off the watch list. Yet the battle isn’t over. Less intervention has meant that the New Taiwan dollar has appreciated a bit. But its currency is now getting back to its 2013 or 2014 levels (against the dollar). Taiwan may want to resume its intervention. Yet with a current account surplus that is still over 13 percent of GDP, it really should tolerate a bit more appreciation. There is one other thing, which is a bit more technical. Taiwan’s life insurers have a ton of foreign assets. And they have added to their foreign asset portfolio at a very impressive clip over the last several years when the New Taiwan dollar was weak. So long as the government intervenes to keep the New Taiwan dollar relative weak, they won’t take large losses on their investments. And as the new Taiwan dollar rises, they take losses. Fair. The risk of such losses helps deter large currency mismatches. But at some point they might need to hedge against the risk of further appreciation. That in turn could accelerate the move… Or put pressure on Taiwan’s central bank to intervene forward. Forward disclosure thus is absolutely critical.

Does China manage its currency against the dollar, against a basket, or to whatever is most convenient at any given point time? Cynics have argued that China seems to peg to the dollar when the dollar is going down and the basket when the dollar is going up. The yuan's moves in March at least raise the question again, even if the signal is relatively weak. The yuan has hewed fairly closely to the dollar over the last few weeks. And in March that meant some modest depreciation against the basket (though the depreciation against the basket was partially reversed in the first week of April when the dollar rose). In other words, had China managed more against a basket, the yuan should have appreciated a bit more against the dollar than it did. Managing the yuan against the dollar is in some ways less risky than managing against a basket, as Chinese residents still seem to focus on the yuan's value against the dollar. Stability against the dollar so far this year—and tighter controls— has contributed to the relative stability in China's headline reserves. It is likely that China is no longer selling all that much foreign exchange in the market, though we still need the settlement data and the PBOC balance sheet data to have a clear picture for March. But I still worry a bit. The risk all along has been that China’s new policy of managing its currency against the basket was masking a policy of managing its currency to depreciate against the basket. From mid-2015 to mid-2016 China used periods of dollar weakness to depreciate against a basket—and thus created the perception of a one-way bet. China needs to make sure such expectations do not reappear, whether by really managing against the dollar—even if that means following the dollar up—or really managing against a basket. One last point: it would be a shame if discussions around currency are defined entirely by the question of “manipulation or not.” It would be a stretch to argue that China is manipulating (there are lots of potential ways to define manipulation, but recently it has been defined as buying foreign exchange in the market to hold your currency down and support a large current account surplus -- and China clearly has selling foreign exchange in the market to support its currency over the past year, and its stimulus has brought its current account surplus down). But the way China manages its currency matters—to the U.S. and to the world. The value of the yuan, more than any other single factor, determines whether it makes sense to locate production in China for U.S. sales, or locate production in the U.S. for sales to China. Of course sectoral barriers matter. But in theory, restrictive Chinese policies in one sector should be offset in part by moves in the currency—fewer Chinese imports should mean a stronger currency, and fewer exports. The sectoral barriers thus matter for particular firms, but the exchange rate matters for the broader economy.* That is why I think China's currency management needs to remain on the bilateral and multilateral agenda and why I disagree with those who argue that other issues (notably the ability of U.S. firms to invest more freely in China) matters more to the U.S. economy than the level of the exchange rate. * As an example, total U.S. beef exports are around $6 billion, with $1.5 billion in exports to Japan and almost $700 million to Hong Kong (call me cynical, but I would not be surprised if some U.S. beef already enters the Chinese market). Even if China started to buy two times as much beef as Japan ($3 billion), it wouldn't really change the overall trade balance by much. U.S. pork exports to China have gotten a lot of attention, but total pork exports to China and Hong Kong still total only around $1 billion a year (total pork exports to the world are $6 billion). Changing the overall trade balance really requires changing the incentives facing a broad set of industries - something that is most easily achieved through a change in the exchange rate.

Martin Wolf’s important column does a wonderful job of illustrating the basic risk China poses to the world: at some point China’s savers could lose confidence in China’s increasingly wild financial system. The resulting outflow of private funds would push China's exchange rate down, and rise to a big current account surplus—even if the vector moving China’s savings onto global markets wasn’t China’s state. History rhymes rather than repeating. And I agree with Martin Wolf’s argument that so long as China saves far too much to invest productively at home, it basically is always struggling with a trade-off between accepting high levels of credit and the resulting inefficiencies at home, or exporting its spare savings to the world. I never have thought that China naturally would rebalance away from both exports and investment. After 2008, it rebalanced away from exports—but toward investment. There always was a risk that could go in reverse too. In one small way, though, I am more optimistic than Martin Wolf. I suspect that China’s regime was under less outflow pressure in 2016 than implied by the (large) fall in reserves, and thus there is more scope for a combination of “credit and controls” (Wolf: "The Chinese authorities are in a trap: either halt credit growth, let investment shrink and generate a recession at home, a huge trade surplus (or both); or keep credit and investment growing, but tighten controls on capital outflows") to buy China a bit of time. Time it needs to use on reforms to bring down China’s high savings rate. All close observers of China know that China has been selling large quantities of reserves over the last 6 or so quarters. The balance of payments data shows a roughly $450 billion loss of reserves in 2016, with significant pressure in q1, q3, and q4. The annualized pace of reserve loss for those three quarters was over $600 billion (q2 was quite calm by contrast). But close examination of the balance of payments indicates that Chinese state actors and other heavily regulated institutions were building up assets abroad even as the PBOC was selling its reserves. In other words, a lot of foreign assets moved from one part of China’s state to another, without ever leaving the state sector. For a some time I have tracked the balance of payments categories dominated by China’s state—and, not coincidentally, categories that experienced rapid growth back in the days when China was trying to hide the true scale of its intervention. One category maps to the foreign assets the state banks hold as part of their regulatory reserve requirement on their deposit base. Another captures the build up of portfolio assets (foreign stocks and bonds) abroad, as most Chinese purchases of foreign debt and equity historically has stemmed from state institutions (the state banks, the CIC, the national pension fund). Even if that isn't totally true now, portfolio outflow continues to take place through pipes the government controls and regulates. These “shadow reserves” rose by over $170 billion in 2016, according to the balance of payments data. The state banks rebuild their foreign currency reserves after depleting them in 2015, and there were large Chinese purchases of both foreign equities and foreign debt. And the overseas loans of Chinese state banks—an outflow that I suspect China could control if it wanted to—rose by $110 billion. As a result much of the foreign exchange that the central bank sold ended up in the hands of other state actors. My broadest measure of true official outflows* shows only $150 billion in net official sales of foreign assets in 2016. Not a small sum to be sure. But not an unmanageable sum. That kind of outflow can easily be financed out of China’s large foreign exchange reserves for a time, or China could more or less bring the financial account into balance by limiting the buildup of foreign assets by the state banks and simultaneously cracking down on outward FDI (over $200 billion in outflows 2016, and largely from state and state-connected companies). In other words, much of the reserve draw was offset by the buildup of other state assets—especially counting the assets state banks and state firms acquired abroad.** That I suspect is why flows suddenly started to balance once China tightened its controls (and likely made it harder for one part of the state to bet against another part of the state). Of course, the renminbi’s relative stability against the dollar also helped—you do earn more on a bank deposit in renminbi than on a bank deposit in dollars. There are of course still private outflows. But the outflows aren’t that much bigger than what China could finance out of its export surplus. If you take the growth in the state banks' overseas assets ("other, other assets" and "other, loans, assets" for the BoP geeks) out of the data, it seems that the pace of outflows actually slowed a bit in 2016. Of course, I do not like simple stories. I think the outflow numbers on the financial account are a bit less scary after you adjust for the financial activity of China's state, but I also think the current account may disguise some real private outflows. Bottom line: the 2016 reserve drain was offset in part by the buildup of assets at the state banks and in other large regulated financial institutions, so it is a bit less scary than it seems—for now. Especially in a context where China’s growth is accelerating, China is on a tightening cycle and, for now, the dollar isn't appreciating in the way many expected. * This measure includes the foreign loans of China's banks (assuming, correctly, that all the major lenders are state-owned), the traces the PBOC's other foreign assets leave in the balance of payments data (other, other, asset -- which reflects the foreign assets the banks hold as part of their regulatory reserve requirement on domestic deposits) and portfolio equity and debt. Historically, the CIC has dominated portfolio equity outflows and the state banks accounted for most portfolio debt outflows. That is a bit less true now, so this arguably over-counts. If that is the case, I would welcome a revision to the Chinese balance-of-payments data that splits out China's holdings of portfolio assets abroad by sector -- consistent with the reporting provided by many other countries. One thing though is clear: portfolio outflows are governed by quotas and are executed through channels the state regulates. They cannot happen without state approval. ** This would explain in part why the data on fx settlement, which includes the state banks as well as the PBOC, shows fewer sales than implied by the change in the PBOC's foreign balance sheet.

Has China's current account surplus disappeared? Should I declare victory and go home? I always have thought a fast-growing, high investment emerging economy should run a current account deficit, not a surplus— Not yet, I would say. It is always dangerous to try to assess China’s trade and balance of payments data for the first quarter, and even more so to opine on the basis of numbers from the first two months. China’s q4 to q1 seasonality is vicious, and the data distortions from the lunar new year are real. But the numbers out now point to some big swings—swings that are worth highlighting. And some real puzzles. I would love to paint a simple picture. But I do not think there is one. The level of chaos in the balance of payments data that the combination of data revisions and real changes seem to have produced is in fact impressive. Some Wild Numbers Consider: In the fourth quarter of 2016, China trade surplus in goods—measured goods, trackable goods—was about $500 billion (annualized). A bit less if you trust the seasonal adjustments. That is a bit off from its peak level of around $600 billion, but still pretty large. The goods surplus incidentally is the easiest to verify in the counter-party data. It generally isn’t off by that much. In the fourth quarter of 2016, China’s current account surplus—the surplus on trade in goods and services (e.g. tourism) net of the balance on interest and dividends from cross border investment—was roughly $50 billion (annualized; the non-annualized surplus in q4 was about $12 billion). That is down from $250-$300 billion in the first part of 2016. The q4 gap between China’s goods balance and its current account was absolutely massive—over $400 billion, annualized. The trade data for the first two months of 2017 points to a fairly substantial fall in the goods trade surplus. The seasonally adjusted numbers point to a fall in the goods surplus under $400 billion (seasonally adjusted)—even after adjusting the data a bit on the assumption that February overstates the swing in the trade balance. And much of this appears to be “real”—not a simple function of changes in iron and oil prices. The average year-over-year increase in import volumes for January and February in the official Chinese data is close to 20 percent (gulp). Of course, these numbers need to be interpreted with appropriate caution: the lunar new year always distorts the numbers in some ways (though combining the January and February data in principle should smooth out the impact) and the March numbers could radically change the picture. If the services deficit and the income deficit (net payments of dividends to foreign companies, relative to the net payment of interest on China’s reserves) are anything like q4, China's current account surplus will have disappeared in the first quarter. The trade surplus is down from q4, and the current account surplus was very small in q4. Broadly speaking outflow pressure looks to have gone way down (see the IIF chart) even as China’s trade surplus has apparently shrunk (or even almost disappeared, after counting services). That of course can happen if a strong expansion pulls in funding from abroad (or in China’s case, convinces Chinese firms who held dollars abroad to profit from an expected depreciation to bring those funds back home). What the heck is really going on? My best guess: revisions to China's method for collecting data on the non-goods balance and real changes in the goods balance combined to produce what appears to be a major swing. I suspect a portion of the fall in the current account in q4 was a result of disguised outflows—not a real change (a portion is also real: there is a reason why emerging economies have rallied…and it isn’t all a function of bets on Trump’s true trade policy). And it does seem like there was a jump in Chinese imports in q1—the rise wasn’t simply due to higher oil prices (oil averaged just over $30 in q1 of 2016, so higher prices for oil and all oil derivatives—petrochemicals for example—are influencing all year-over-year comparisons). But I want to see more data before I draw any strong conclusions—January and February really need to be joined with March to get a better picture of the overall trend. 20 percent year-over-year growth in import volumes seems a bit too high to be sustained, even for a booming China. These arguments aren’t very "Zero Hedge-y". And they sort of go in offsetting directions. Many are informed by a sense that China’s current account surplus hasn’t really gone from $250 billion to $0-$50 billion in a couple of quarters, even if it likely has gone down. A $200 billion current account swing should be big enough to have some really significant effects globally. I see some knock-on effects (the emerging market rally for example), but, for now, not big enough ones for me to think the full swing is real. At this stage, all the arguments are more working hypotheses than anything else. Something is going on, it just isn’t completely clear what. The Tourist Trap To start, China apparently revised its tourism data (yet again). The tourism deficit is about 2 percent of GDP—so the estimated tourism deficit matters for the overall balance of payments. China’s tourism exports—spending by foreign tourists in China—got revised down (after being revised up in the last revision). The flat trajectory of spending by foreign tourists in China is consistent with arrival numbers. But the revisions are a substantial change from what was in the data previously, even if the last rounds of data largely undid the effect of the previous revisions (if that makes sense). China’s tourism imports jumped in q4-2016, helping bring the current account down (in the seasonally adjusted numbers). And this possibly reflects the impact of the latest revisions—which didn’t change the trajectory of tourism imports (spending by Chinese residents abroad) nearly as much as they changed the trajectory of China’s tourism exports. The big 2014 jump in reported tourism imports remains. That corresponds to the introduction of a new methodology of collecting the data. Take out the q1-2014 jump, and the growth in tourism is more or less in line with the data on the number of Chinese residents travelling abroad. The second jump—in q4-2016—shows up more obviously in the seasonally adjusted numbers (seasonal adjustment by the way is sensible for tourism: tourism, like the harvest, has a pronounced seasonality). Tourism spending is estimated to have increased by 20 percent year-over-year in q4. Which is possible, but it seems a bit out of line with changes in the number of Chinese residents travelling—I am not aware of a big jump in outbound tourism in q4. (The data for h1 suggests the number of outbound tourists was flat to up slightly). The problem with the “new” tourism data is actually simple: the new methodology almost certainly overcounts actual spending abroad by Chinese tourists, as it counts purchases of assets abroad as tourism imports. Take the Bloomberg story on Chinese residents purchasing expensive art abroad for example. I guess that probably should be counted as a goods imports. But no doubt a portion of the tourism outflow is going into financial assets abroad, and thus really should be in the financial account. Remember that the Chinese tourism import numbers no longer match up to the data reported on tourism exports in the biggest destinations—notably Hong Kong. Either the Chinese numbers are off, or the numbers reported by countries receiving large numbers of Chinese tourists are off. Suffice to say that since 2013 the number of outbound tourists has increased from about 100 million to around 125 million. And tourism imports in the balance of payments have gone from about $125 billion to about $260 billion (in the revised data). The methodological change in how China reports its tourism data (introduced in 2015, but with an impact on the 2014 numbers) has had a huge impact. Disguised Outflows Certainly the q4 gap between the goods surplus and the current account surplus is so big that it seems reasonable to think that some financial outflows are now disguised in the current account. Look at a plot of the goods balance against the current account balance. The goods balance came down in the fourth quarter of 2016—for real (see this post). China’s manufacturing surplus has slipped a bit in 2016, after rising in 2014. But most of the change in the current account reflects a growing gap between the reported current account and the reported goods balance. And that is driven by “tourism” outflows, not by the rest of the data. The current income deficit isn’t out of line with the basic trend. With an annualized tourism deficit that is now close to $250 billion, the question of how much of China's tourism deficit is real, and how much of it reflects disguised outflows matters for the global balance of payments, and for assessments of currency misalignment.* The Recent Goods Data In some ways the fall in China’s goods surplus in q1—or really the first two months of q1—is more interesting than the fall in the q4 current account. Obviously q1 is heavily influenced by the lunar new year, and the q1 data tends to show pronounced seasonality. And equally clearly some of the rise in the deficit is from higher commodity prices, and thus fully expected. But the year-over-year data also shows a really large rise in import volumes—combined January/February import volumes were up 20 percent in the official data. Combined January and February export volumes were up about 5 percent. That is a huge gap. If it is confirmed in March, it will suggest that China's overall surplus really is falling on the back of China’s stimulus. I have long thought the slowdown in China's import volumes back in q1-2015 sent an important signal of China’s economic cycle. And right now the trade data points in the opposite direction. March will help establish just how strong that acceleration is. Of course the acceleration in imports is probably more pronounced than the actual Chinese cycle, as imports tend to reflect trends in investment more than trends in services and are heavily influenced by demand for commodities. And, if China is pulling back on its stimulus, the pace of import growth should moderate. In Martin Wolf's terms, China looks to have escaped the trap of high savings through a combination of credit stimulus and controls in the first quarter, but sustaining the current equilibrium may well require sustained stimulus. And the sharp rise in imports also raises the possibility that goods imports have been overstated as a way of moving money out of China in an era marked by stronger enforcement of controls. Sort of like the travel data. That said, there is no question that China’s current account has suddenly become quite volatile. In part that is because a high fraction of China’s goods imports are commodities (the share of manufactured imports has slid as a share of GDP) and commodity prices are volatile. In part it is because of the explosive growth in reported tourism imports, China's propensity to revise the tourism numbers, and the possibility that the current account is now influenced by disguised financial outflows. And in part it is because of new volatility in the import volume numbers. Twenty percent year-over-year increases in volumes aren't common any more, especially when the volume growth isn't a reflection of exports. Bottom line: A roughly $200 billion (greater than 1.5 percent of China's GDP) swing in the current account over the course of a couple of quarters doesn’t happen that often, absent a true oil shock or a financial crisis. It demands both attention, and a bit of skepticism. * The current account balance, not the the trade balance, is the variable used in the IMF’s multilateral exchange rate surveillance, and in the United States' foreign exchange report. I at times do wonder if that influences decisions about how to revise China’s balance of payments methodology. The key changes have all had the effect of reducing China’s current account surplus relative to its trade surplus (they are also defensible on technical grounds).

The proxies for Chinese reserve sales show very modest sales in February. Foreign exchange settlement (which includes the state banks) shows $10 billion in sales, and only $2 billion counting forwards. The PBOC’s balance sheet shows similar changes—foreign reserves fell by $8.5 billion and foreign assets fell by $10.6 billion. No wonder the (fx) market is no longer focused on China. The fall off in foreign exchange sales is particularly impressive given that China didn’t have its usual trade surplus in February, for seasonal reasons (China’s trade often swings into deficit during the lunar new year). Modest reserve sales alongside a monthly trade deficit imply that the pace of capital outflows fell. The only analytical problem is that the fall in pressure on the renminbi is a bit over-determined. Controls on outflows were tightened. For real, it seems. That likely helped. And the yuan was stable against the dollar, broadly speaking. There continues to be a correlation between movements in the yuan and the pace of outflows. In other words, to stabilize outflows, it really helps if China at least keeps the yuan stable against the dollar. Any depreciation (against the dollar) tends to generate expectations of more depreciation. Chinese firms with dollars want to hold on to them. And so on. Appreciation against the dollar by contrast tends to encourage firms with dollars to sell their dollars and bring their money home. The interest rate on the yuan is higher than on the dollar remember; holding dollars can be costly. The data for March won’t be out for several more weeks. Given the dollar sold off a bit in March, allowing the yuan to rise a bit against the dollar within the broad context of managing against a basket—there is no strong reason though to think pressure on the yuan has returned. One implication: the money market tightening in China likely was done for domestic reasons, not as a by-product of an effort to support the currency. I though do wish that China had been a bit symmetrical in its management against the basket, and had not (once again) used the dollar’s weakness as an opportunity to let the yuan slide (a bit) against the CFETS basket (see the chart above). China needs to add to the two way risk around the yuan. That will not happen if the big players in China’s foreign exchange market think the yuan will be stable against the dollar when the dollar is going down, and will depreciate against the dollar when the dollar is going up (and thus grind down against the basket). So in my view, China either missed an opportunity to let the yuan rise a bit more against the dollar and add to two-way risk, or signaled that it has reverted to managing the yuan against the dollar once again.

It should go almost without saying that China’s ability to maintain its current exchange rate regime matters. The yuan has been more less stable against the CFETS basket since last July. If the current peg breaks, China will struggle to avoid a major overshoot of its exchange rate. Christopher Balding recently has argued that the fall in the dollar (on say the Fed’s dollar index) in 2017 is more or less the same as the yuan’s depreciation against the basket—which would make China’s exchange rate regime now more a pure peg against the dollar rather than a true basket peg. Hence the lack of movement against the dollar in past few weeks. Maybe. I though am inclined to think that the yuan’s depreciation against the basket this year just undid the upward drift against the basket that came when the dollar appreciated last year, and China is still aiming to keep the currency more or less stable against the basket, not stable against the dollar. Time will tell. Right now, the exact nature of China’s peg now matters less than China’s ability to maintain some kind of peg, and thus to avoid a sharp depreciation. If there is a depreciation, either the world will absorb a new wave of Chinese exports—a 10 percent real effective exchange rate depreciation should raise China’s real trade balance by about 1.5 percent of China’s GDP, or by roughly $180 billion (using this IMF study as a baseline for the estimate, it is summarized here)—or a wave of protectionist action will limit China’s export response, and in the process threaten the global trading rules.* Neither is a good outcome. The data for the next few months will be critical. China does seem to have tightened its outflow controls—despite the official denials. FDI outflows certainly have slowed. Hopefully the screws will be placed on a few other categories of outflows—there are plenty of categories in the balance of payments that show a buildup of foreign assets that, in my view, should be controllable. The rise in Chinese bank loans to the world, for example. I also still think the yuan’s movement against the dollar is an important driver of outflows, so a sustained period of stability in the yuan’s exchange rate versus dollar—whether because the broad dollar is falling and that means the yuan needs to rise to maintain a basket peg, or simply because China starts to prioritize stability versus the dollar—should lead to a reduction in pressure.** Finally China does seem to be tightening its domestic policy, at least a bit. Higher money market rates should also help support the yuan. The proxies for intervention for January do suggest some reduction in pressure in January—though pressure by no means entirely went away. The FX settlement data, adjusted for reported forwards, shows net sales by the banking system of just under $20 billion. That is down from $36 billion in November, and $54 billion in December. The PBOC’s balance sheet data shows a $30 billion fall in foreign reserves, and a $60 billion fall in foreign assets. That can be compared to an average monthly fall—on both measures—of around $45 billion in the fourth quarter of 2016. I typically prefer foreign assets to foreign reserves, because foreign assets captures the foreign exchange the banks hold at the PBOC as part of their regulatory reserve requirement—and in the past, changes in the banks required reserves have been a tool that the central bank has used for shadow intervention. However, the January fall looks to be something else. In the past, changes in the PBOC’s foreign assets—at least those that mapped to backdoor intervention—generally have been correlated with changes in the settlement data (this makes sense, the PBoC’s other foreign assets are in part the banks required reserves, and the settlement data aggregates the central bank and the state banks). One story is that the PBOC has changed how it accounts for its contribution to the IMF. But until that story is confirmed, the fall in other foreign assets has to be part of the calculus. While the gap between the settlement number and the change in the PBOC’s balance sheet is particularly large in January, a smaller gap has been present for a while. The PBOC’s balance sheet data thus implies slightly larger outflows than the settlement data. Bottom line: The balance of evidence suggests a moderation in pressure on China’s exchange rate regime in January—but a moderation in pressure isn’t the same as an end to the pressure. Unlike some, I do not think China’s fundamentals require a depreciation. The current account remains in surplus, China’s export market share has held up well—and China is (once again) growing faster than most of the world. But sustaining the current basket peg will be hard if outflows do not moderate. * I recognize that a rise in protectionism is a possible outcome even if China doesn’t move ** If stability or appreciation against the dollar doesn’t materially reduce reserve loss over a several month period, I would need to reevaluate my priors

$2.7 trillion is well over 3 times China’s short-term external debt (around $800 billion per the IMF). It is roughly two times China’s external debt ($1.4 trillion, counting over $200 billion in intra-company loans). It is enough to cover well over 12 months of goods and services imports (total imports in 2016 were around $2 trillion).* There are two good reasons why a country might need more reserves than it has maturing external debt. The first is that it has an ongoing current account deficit. A country arguably should hold reserves to survive one year without any external financing – the sum of the current account and short-term external debt. The other is that a country has lots of domestic foreign currency deposits. Neither applies to China. China’s runs a $200 to $300 billion current account surplus, so its one year external financing need is now around $500 billion. China has a relatively modest $250-$300 billion in foreign currency sight deposits, and just under $600 billion in domestic foreign currency deposits (to put that in context, it is about 5 percent of GDP). It would take just over a trillion in reserves to cover China’s external (short-term debt) and internal (domestic fx sight deposits) fx liquidity need. The $2.7 trillion number (or $2.6 trillion) stems from the initial application of the IMF’s new (and in my view flawed) reserve metric to China. The IMF’s metric revived M2 to reserves as an important indicator of reserve adequacy, and China is off the charts on this indicator (China has very little external debt, but a very large domestic deposit base – so a composite indicator that includes the domestic deposit base gets a very different result than metrics that focus on external debt). In the composite indicator, emerging economies with a fixed exchange rate and an open capital account need to hold 10% of M2 in reserves. That alone is about over 20% of China’s GDP – as China’s M2 to GDP ratio is a bit over 200%. For China, the weighted contributions from short-term debt, “exports” (the IMF uses exports rather than imports) and long-term external liabilities are trivial. For China, the entire reserve need more or less comes from one of the four variables in the IMF’s composite indicator (more here). But that calculation is now outdated. The IMF has refined its metric to give more weight to the presence of capital controls, and China has tightened its controls. Assuming that there are capital controls, the IMF metric indicates that China would be fine with $1.8 trillion in reserves (though that sum rises over the course of 2017, thanks to the ongoing growth in M2 as a share of GDP).** I am not a fan of the even the updated new metric. I am not a big fan of composite metrics in general. And if you are going to use a composite metric, I think the composite metric should put more weight on foreign currency deposits than domestic currency deposits, while the IMF’s metric typically weights all domestic deposits at 5%. The IMF’s metric thus ignores one of the key insights of balance sheet analysis. No matter. The world would be in a better place if there was a broad recognition that China can burn through another $1 trillion in reserves and, with $2 trillion still in reserves, be above every nearly all metrics of reserve adequacy. This isn’t to downplay the scale of China’s reserve loss. The BoP data shows a reserve outflow of a bit more than $440 billion in 2016. That is significant. But I suspect that much of the outflow that led to the fall actually was the state actors. The build up of foreign assets in the banks, loans from the state banks to the world and easily controllable portfolio outflows from large institutions likely accounted for about $250 billion of the total fall in reserves. This leaves the true fall in the total foreign assets of China’s state sector at more like $200 billion. Still big, but not quite as big. Set that debate aside though. The reason for concern about China is the rapid pace of the reserve decline, not the risk that China is about to run out of reserves. China might conclude it doesn’t want to continue to finance outflows with reserves, in which case its currency would depreciate until the trade surplus was large enough to finance the outflow. One final point. The IMF’s initial metric implied China needed far more reserves, relative to its GDP, than other emerging markets economies – more than Brazil, more than Turkey, more than Russia, more than Ukraine (see this post, or play with the IMF’s data tool***). That never made sense to me, even though China does have a more managed exchange rate than most. Domestic capital flight can force a country off a peg, no doubt – but the negative macroeconomic impact of a depreciation is proportionate to the amount of foreign currency debt in the economy, not the size of the domestic deposit base. *  The IMF metric uses exports rather than the traditional imports, but China is just fine using exports too. China’s export to GDP ratio has come down after the global financial crisis. ** In the absence of capital controls, the IMF’s latest reserve metric spreadsheet suggests China now needs more than $2.7 trillion using the 10% of M2 weighting, incidentally. The estimated 2017 reserve need is -- gulp -- $3.3 trillion. One way of interpreting that is that China lacks the reserves to open its financial account if it wants to continue to manage its exchange rate. Another way of interpreting it is that the metric doesn’t really work for an M2 to GDP outlier like China. *** The IMF’s reserve data interface is now quite useful. And all the underlying data is available in a convenient spreadsheet. That makes it easy to check out the contribution of various components. I would be thrilled if the IMF added the foreign currency/ domestic currency split of M2 so that those of us who are interested in the foreign currency deposit base would have a new resource for cross country comparison.

Fred Bergsten of the Peterson Institute has long argued for target zones around the major exchange rates. In the past twenty years, no G-3 economy has followed Bergsten’s policy advice.* The world’s biggest advanced economies have wanted to maintain monetary policy independence, and—Japan perhaps excepted, at least during certain periods—they haven’t viewed foreign exchange intervention as an independent policy tool. China formally has a band around its daily fix, but its exchange rate is still more of a peg (now a basket peg, without any obvious directional crawl over the last 7 months) than anything else. I though increasingly think that Korea’s exchange rate management could be described as a target zone of sorts. It buys dollars and sells won when it thinks the won is too strong (recently, too strong has been less than 1100 won per dollar). And it sells dollars and buys won when it thinks the won is too weak (recently too weak has been above 1200 won per dollar). As a result, Korea intervened heavily to cap won strength in q3 2016. Counting the change in its forward position, government deposits, and government purchases of foreign debt securities, its total foreign exchange purchases in q3 were roughly $12 billion (about 3.5 percent of GDP).** It then turned around and sold dollars to limit won weakness in q4 2016. Using the same metric, total sales reached $8 billion, or around 2.5 percent of GDP (counting the change in the forward book is important here). And it presumably stayed mostly on the sidelines in January, as the won moved within Korea’s zone of intervention (the Koreans also have been hinting that they may need to intervene a bit less to stay off Trump’s radar. See Christine Kim of Reuters. Worth watching.) Korea’s target zone isn’t declared. It has to be inferred from watching the levels where Korea intervenes. And the level where the Koreans intervene has shifted over time. For most of the post-crisis period, they intervened at 1050 against the dollar—and briefly they allowed the won to appreciate all the way to 1000 before intervening. But over the past couple of years Korea’s intervention to limit won appreciation has typically come at around the 1100 mark.*** And this raises an important question for both the IMF and the U.S. Treasury in their exchange rate surveillance. Does the exchange rate where intervention occurs matter, or just the amount? Korea’s total intervention over the last 12 months—even with my adjustments for forwards and the debt purchases by the government—hasn’t been significant. After q4, the government on net looks to have sold about $5 billion over the course of the year. It on net bought in the second half of 2016. But its net buying in the second hold of the year certainly doesn’t cross the 2 percent of GDP thresh-hold the U.S. Treasury set out a year ago. But looking only at total intervention over the course of a turbulent year, in my view, also misses a lot of the story. Korea bought and sold, buying a bit more than 3 percent of GDP (annualized) in q3. But its buying and selling in my view raises the question of whether it is buying and selling at the right levels. And given Korea’s ongoing trade and current account surpluses, I would argue the answer to that question is still "no." * Edited after posting. Fred Bergsten has indicated the post-Louvre reference rates were a target zone of sorts. I confess that I wasn’t thinking back to the 1980s when I wrote this. ** I have left out the National Pension Service’s ongoing purchases of foreign equities, which totals—based on the balance of payments data—around $12 billion (roughly 1 percent of GDP). That ongoing flow certainly raises Korea’s current account surplus (it is financed out of tax withholdings, as the social security system takes in far more in contributions than it pays out) and structurally weakens the won. But it isn’t classically considered intervention. *** Before the summer of 2014, Korea typically intervened at the 1050 mark. It briefly allowed the won to rise toward 1000 in the summer of 2014. But after the dollar strengthened in late 2014, Korea has not allowed the won to rise above 1075 on a sustained basis.

   

The pace of decline in China’s foreign reserves matters. Not because China is about to run out. But rather because China will at some point decide that it doesn’t want to continue to prioritize “stability” (against a basket) and will instead prioritize the preservation of its reserves, and let the yuan adjust down. Significant voices inside China are already making that argument. And I fear that if the yuan floats down, it will stay down. China will want to rebuild reserves, and—if exports respond to the weak yuan—(re)discover the joys of export-led growth. Relying on exports is easier than fighting the finance ministry’s opposition to a more expansive (on-budget) fiscal policy, or seriously expanding the provision of social insurance to bring down China’s savings. I thus disagree with those who argue that the “China” shock is over. It depends a bit on the exchange rate. China’s exports of apparel and shoes have probably peaked. But China’s exports of a range of machinery and capital goods continue to remain strong—and at a weaker exchange rate, China could supply more of the components that go into our electronic devices, and export far more auto parts, construction equipment parts, engines, generators, and even finished autos than it does now. “Mechanical” engineering writ large continues to be a significant part of the U.S. economy, and even more so the European economy. One of the main indicators—PBOC balance sheet reserves—that I follow for tracking China’s reserve sales is now out for December, and it points to around $45 billion in sales. I prefer to look at all the foreign assets the PBOC reports on its balance sheet rather than just its reported foreign exchange reserves. That variable was down $43 billion in December, and $133 billion for q4. Actual foreign exchange reserves fell by a bit more—$46 billion in December and $141 billion in q4. The difference between foreign exchange reserves and all of the PBOC’s foreign assets is primarily the foreign exchange the banks hold at the PBOC as a result of their reserve requirement. The loss of reserves in December was a bit smaller than in November. But only just. The average monthly fall in q4 was over $40 billion. That is a pace that is ultimately unsustainable. I think China would be fine with $2 trillion in reserves, given how little foreign debt it holds. Others say $2.5 trillion. If reserves are falling by a steady $40 billion a month/$500 billion year, it is only a matter of time before China hits its limit. With China, it may be a long time though… However, there are two reasons why I am not yet convinced that it is only a matter of time before outflows overwhelm the PBOC’s reserves and other exchange rate defenses. First, reserve loss and thus the scale of outflows remains correlated with movements in the yuan against the dollar. In quarters with significant yuan depreciation against the dollar, reserve loss—using the PBOC’s foreign assets as the measure—has varied between $100 billion and $225 billion (q3). In the two quarters during the past year when the yuan was stable or appreciating against the dollar (q2 of 2015 and q2 of 2016) the quarterly reserve loss was very modest—under $20 billion. So if the PBOC was serious—and that might mean losing a bit of face by moving away from the basket peg—and let the yuan strengthen against the dollar, my guess is that outflow pressures would fall significantly. It would help if the PBOC allowed a bit (more) upward drift against the basket—and thus reinforced expectations of two-way risk against the dollar. Remember that the PBOC took advantage of dollar weakness in the first half of 2016 to reset the yuan’s level against the basket. I understand why, but that reset had a price—it reinforced expectations that the yuan only will move one way against the dollar. The second is that a lot of the outflows so far in 2016 have come through channels that I think the State Administration for Foreign Exchange (SAFE) can effectively control, if it was determined to do so. A surprisingly large share of the outflows last year come from balance of payments categories that I think the authorities can control, and can control without too much administrative difficulty, as they ultimately require supervising a relatively small number of accounts at a manageable number of large state institutions (of course there are political difficulties here, but the administrative complexity should be lower). A bit of balance of payments math: China runs a goods surplus of around $500 billion (in BoP terms) annually. And—even with the fall off in FDI inflows—should get another $100-$150 billion in FDI ($250 billion would have been more typical a few years back). So China has over $600 billion/6 percent of GDP in inflows from FDI inflows and the goods trade to finance its services imports (which likely include a lot of hidden outflows through the tourism side), its FDI outflows, and non-FDI outflows, without having to dip into its reserves. Of course, China has dipped into its reserves in 2016. Based on the PBOC balance sheet data, reserve outflows in the balance of payments for 2016 should be about $450 billion (I am assuming a $150 billion fall in reserves in the q4 balance of payments). But I can count outflows of roughly equal size that seem to me relatively easy for the PBOC—really SAFE—to control if they really want to. In 2016, the banks have been adding to their foreign exchange reserves (other foreign assets) at the PBOC at a roughly $25 billion a quarter pace. That slowed in q4, so let’s call it about $75 billion a year in outflows. That easily could be put to an end; the share of the banks regulatory reserves held in foreign exchange is totally determined by the PBOC. In 2016, the build-up of portfolio debt and equity assets abroad—almost certainly by a few state institutions or major financial institutions that the state regulates—has been a bit under $25 billion a quarter. Call it $100 billion a year. China knows how to put a stop to these flows; outflows in these line items were essentially zero from 2010 to 2014. In 2016, overseas lending—long-term loans made by the Chinese state banks to the rest of the world—have averaged a bit over $25 billion a quarter, or $100 billion annualized. These all come from the big state commercial and policy banks. They could be slowed with a bit of regulatory supervision. And FDI outflows have averaged about $60 billion a quarter in the first three quarters (roughly $250 billion annualized). A more normal number—judging from the numbers seen before the devaluation—would be maybe $25 billion a quarter ($100 billion a year). Reducing outflows there back to the norm—as seems likely to happen—might reduce total outflows by $150 billion. Sum up these line items, and that takes away $425 billion in outflows—a sum almost equal to the projected fall in reserves in the balance of payments. And I suspect that if China’s reserves fell by $25 billion a year, no one would care that much (even if such a number implies an ongoing outflow of between $400 and $500 billion, depending on your view of how much of the services deficit is “real”).

   

I want to step back a bit from the rather extraordinary moves in the offshore yuan market over the past few days. It seems quite clear that China’s authorities felt the need to signal that the yuan isn’t currently a one way bet against the dollar. And stepping back in this case means taking a deep dive into the details of the balance of payments data -- details that come out with a quarter lag, and thus provide information that is stale from the point of view of a forward-looking market. A lot, and I mean a lot, changed in the fourth quarter. I generally like it when China’s data series line up. Line up with each other. And, when possible, when China’s data also lines up with data reported by China’s trading partners. So I have been bothered for some time by the large discrepancy between the fall in China’s foreign exchange reserves (as reported on the PBOC’s balance sheet, $108 billion in the third quarter) and the much smaller net sales of foreign exchange by China’s banks (as reported in the FX settlement data, $50 billion in the third quarter without adjusting for the forwards reported in the settlement data, $63 billion with the forward adjustment). Fx settlement includes all the banks, not just the central bank. Historically, though, it has been very correlated with overall reserves. The initial balance of payments (BoP) data for the third quarter showed large reserves sales ($136 billion), sales on a scale that was consistent with the PBOC balance sheet numbers. The BoP reserves sales thus seemed to suggest a big pickup in capital outflows in the third quarter. However, the detailed balance of payments data suggests that the signal from the FX settlement data may be more accurate. Much of the q3 fall in China’s reserves seems tobe explained by the buildup of foreign assets by other state controlled financial institutions, not “private” capital outflows. I see a likely increase of around $85 billion in the foreign assets of state institutions other than the PBOC in q3. Which implies that the actual fall in “official” assets might be as low as $50 billion in the third quarter, and estimated private outflows (estimated as the difference between net inflows from the current account and net FDI flows and the buildup/sale of official assets) might have been about $100 billion. The ongoing build-up of foreign assets by the state banks and other state institutions also implies that the net sale of foreign assets by China’s state might have been as low as $200 billion in the four quarters from q4 2015 to q3 2016, with $400 billion in reserve sales offset in part by a rise of around $200 billion in the foreign assets of China’s other state financial institutions. Two hundred billion dollars is a large sum -- but one that also suggests that depreciation pressures earlier this year were significantly smaller (when official sales are scaled relative to China’s GDP) than the appreciation pressure back in 2006, 2007 and 2008 (and, for that matter, 2010). (In the chart above estimated reserves is a bit mislabeled; it should be either "augmented reserves" or "estimated official outflows") To be clear, my adjustments change the timing of the sales (and thus the timing of the peak outflow pressure) more than the cumulative magnitude of sales. They imply that state financial institutions sold a lot of foreign assets in the third quarter of 2015 to help the PBOC support the currency, and then rebuilt their foreign asset position over the next few quarters -- most obviously through a fall and then a rise in the state banks holdings of foreign exchange as part of their required reserves (reported as "other foreign assets" on the PBOC’s balance sheet).   And I also have little doubt that outflow pressure increased in the fourth quarter of 2016. The modest sum in the trailing four quarter sum through q3 is thus in part a function of the fact that two quarters with large sales are just outside the covered time period. Several other caveats are also needed. First, I am trying to understand the outflows in the third quarter (July through September). The third quarter is now ancient history. The yuan was relatively stable against the dollar in most of the third quarter. The yuan decidedly was not stable against the dollar in q4 and outflow pressures seem to have increased. The PBOC tightened its controls in the fourth quarter for a reason. Second, even if I am right and "true" private outflows are smaller than implied by reserve sales in the balance of payments in the third quarter —the obvious measure of “hot” outflows from China is either the difference between reserves (sales/purchases) and the current account, or the difference between reserve sales and the combined surplus from the current account and net FDI flows—there is no question that outflows were larger in the third quarter of 2016 than in the second quarter of 2016. The question is one of magnitude, not of direction. Third, to keep it simple, my estimate of "private" financial outflows in the first chart doesn’t include outflows that are disguised in the current account. It thus likely under-states actual capital outflows, as I clearly think there are financial outflows embedded in the current account. I tried to keep the chart relatively simple, and the adjustments relatively transparent -- so I didn’t adjust for hidden outflows.** Fourth, these estimates hinge on a set of educated guesses about who appears where in the balance of payments data, as China doesn’t release the kind of details that would allow outside observers to know for sure which set of institutions inside China’s financial system built up their foreign assets. I want to emphasize that these estimates are subject to change as my understanding of what drives different categories in the balance of payments changes. So why precisely do I suspect official (or quasi-official) asset growth provided a significant offset to reserve sales in the third quarter? A) The balance of payments data shows a $16 billion increase in the line item that corresponds to the rise in the PBOC’s other foreign assets -- the technical term for the foreign exchange the banks are required to hold as part of the regulatory reserve requirement. B) There was a big rise in lending by the state banks to the rest of the world ("loans" by the banks to the world rose by $37 billion in q3). I have always assumed that the term “state commercial bank” has real meaning in China, and that such lending could be stopped if the government really wanted it to stop. It was more or less non-existent before 2011, and much of the rise since then clearly comes from the China Development Bank. C) The balance of payments data shows large purchases of foreign portfolio debt and foreign portfolio equity (just over $30 billion in q3). This is a bit puzzling because there was no reported change in the Qualified Domestic Institutional Investor (QDII) quotas. So it wasn’t, I assume, from small private investors. And it wasn’t totally hidden either. After all, it shows up in the official data as an outflow, and also as a rise in foreign assets in the international investment position data. My best guess is that this is coming from some kind of state institutions —be it the social security fund, the China Investment Corporation (CIC), the state banks or even perhaps the life insurers. As a result, my estimate of the total stock of foreign assets under China’s state control fell by significantly less than the PBOC’s reserves in the third quarter.** All this implies that one arm of the state is betting against another, with state financial institutions building up assets abroad even as the PBOC runs down its reserves. And it thus raises a simple question: why would the government want to draw down its central bank reserves to allow the state commercial banks and other state financial institutions to raise their foreign assets? My gut is that some folks at the PBOC started asking similar questions over the last year, which could in part explain the tightening of controls toward the end of this year—cracking down on FDI outflows sent a signal that the "go out" policy no longer dominates all other considerations. Chinese policy coordination isn’t always perfect. Finally, my gut also says that while the pressure in q3 wasn’t as severe as the fall in reserves suggests, the fourth quarter is different. The squeeze now underway in the offshore yuan market suggests that the PBOC is working hard to discourage further outflows and to fight expectations of further depreciation. * Technically, I am assuming that "other, other, assets", "other, loans, assets", "portfolio debt, assets", "portfolio equity, assets" are all driven by state controlled institutions, and not making further adjustments The portfolio flows data could be adjusted for QDII, but I think it is clear that "portfolio debt" historically has been driven by the state banks (05 to 06 purchases, then sales) and portfolio equity has been in part driven by the CIC given the timing, and the recent run-up isn’t a function of QDII. "Other, other" is the foreign exchange the banks hold at the PBOC as part of their required reserves, other, loans seems historically to have been driven by the CDB. Of course the recent rise could come from different channels, but it almost certainly comes from the large state banks. ** This is all based on "on balance sheet assets," it doesn’t attempt to adjust for China’s rumored forward book. See this post for more on the forwards (with apologies for the level of technical detail).