Trey McArver

Director, China Research
+44 20 7246 7894

How can we help?

To learn more about our services get in touch on +44 20 7246 7800 or

Recent China research

* China Watch: Strong global growth a tailwind for China's deleveraging, Michelle Lam, Trey McArver, Larry Brainard, 14 Dec 2017
* China Watch: Outlook for 2018, Larry Brainard, Trey McArver, Michelle Lam, 7 Dec 2017
* China Perspective: North Korea scores, US stuck, Jonathan Fenby, 1 Dec 2017
* China Watch: The profit cycle – still some way to go, Michelle Lam, Trey McArver, Larry Brainard, 30 Nov 2017
* China Watch: China’s road to an automotive superpower, Michelle Lam, Trey McArver, Larry Brainard, Kingsmill Bond, 23 Nov 2017

See all our China research >>

The real anatomy of Chinese capital outflows


Capital outflows from China accelerated appreciably in 2015. We estimate the net total at US$870 billion, slightly smaller than the often quoted figure of US$1 trillion but still very significant at nearly 9 per cent of GDP. We expect this process to slow down in 2016, and flows may even reverse if expectations of further dollar weakness take hold. Meanwhile direct investment remains robust, indicating that business sentiment on the domestic economy remains positive. But capital outflows are affecting the real economy by impairing liquidity conditions, and China’s monetary authorities have only limited measures with which to address liquidity issues in view of their desire to maintain a stable exchange rate.

Key judgments

  • China’s capital outflows are primarily the result of Chinese corporations seeking to increase dollar assets (and to decrease dollar liabilities) in response to a weakening RMB and expectations of further weakness.
  • Thus the majority of China’s capital outflows do not represent capital flight, and the drawdown in China’s FX reserves can be seen as a shift of dollars from the government balance sheet to the corporate balance sheet.
  • Our base-case scenario is that capital outflows will slow down, or even reverse, owing to tighter capital controls, the end of asset rebalancing by corporates, the weakening of the dollar and more positive sentiment about the Chinese economy.
  • Downside risk still exists; the worst-case scenario is that outflows from households accelerate, creating self-reinforcing depreciation pressures that could lead to a currency crisis.
  • The long-term goal remains a freer exchange rate and a more open capital account, but the government’s bias towards control means that these reforms remain on hold until volatility in the currency is brought under control.


This note examines the capital outflow situation in China, which has been clouded by misunderstanding and misleading reporting. We aim to answer the following questions that must be clarified before any serious calculation can be made of the extent of the outflows and their long-range impact:

  • What is driving the outflows?
  • What are the implications for the Chinese economy?
  • How is the government addressing the issue?
  • How big are the total outflows?
  • What lies ahead?

Where has all the money gone?

We discuss how we arrived at the figure of US$870 billion net capital outflows in 2015 later in the note, but we think that understanding where this money is going and what it is being used for is much more important than how large the outflows are. Not all capital outflows are bad, and the implications for the economy and the currency can be understood only if we disaggregate the outflows so as to understand what is driving them. On the whole, we find that very little of the outflow total represents capital flight. It results instead from the shifting of corporate (and, to a lesser degree, household) assets into dollars in anticipation of expected RMB depreciation. Most of these “outflows” are thus still held by Chinese entities and will “come back” when the RMB is no longer expected to depreciate.

Capital outflows are not, in and of themselves, definitively indicative of capital flight or an unhealthy economy. In the light of China’s large trade surplus, we would expect a negative financial account balance, similar to what is seen in other countries with large trade surpluses such as Germany and Japan. Although much larger in absolute terms, China’s outflows as a percentage of GDP (8.5 per cent) were smaller than that of Germany in 2013 (8.9 per cent).

But whereas Germany’s outflows more or less mirrored its current account surplus, China’s are three times its current account surplus, implying that there is more going on here than the recycling of trade-related inflows. However, there is strong evidence that a sizeable proportion of the outflows represent benign forces, such as corporates practising prudent asset management or paying off or hedging dollar debt. There has also been more overseas lending by Chinese financial institutions. These are healthy flows, and are not capital flight.

Chart 1: Annual net capital flows, 2003-15

Because the RMB is an international currency, we cannot get a full picture of China’s capital flows from looking at its balance of payments (BOP) data. Nonetheless it remains the best place to start for understanding China’ capital flows. As much of the relevant Q4 data has yet to be released (and will not be available until March), we will focus on the outflows through Q3. This is an important omission in that roughly one-third of outflows took place in Q4. But by analysing the composition of outflows through Q3, we do get a picture of what is driving them; and this will help us to make an educated guess about what happened in Q4.

Looking at the data, we see five categories that have an appreciable effect on capital outflows: portfolio investment, currency and deposits, loans, trade and credit advances, and errors and omissions (see Chart 2 and Table 1 below). Together these five components made up US$614 billion of outflows through Q3, although total net outflows were smaller because of other offsetting factors such as a positive balance on accounts receivable.

Chart 2: Contributors to capital outflows

Table 1: Contributors to outflows as percentage of total, Q1-Q3/15

We can see in Chart 2 above that portfolio investment flows turned negative in 2015. Historically foreign investors have sought ever-greater access to China’s securities markets, and portfolio flows had been positive since 2006. The balance of flows was relatively small at -US$ 41 billion, but the more than US$120 billion change in direction was the biggest contributor to the growth of capital outflows through Q3/15, accounting for 40 per cent of the increase in these five categories (see Chart 3 below).

Chart 3: Proportional contribution to change in outflows, 2014-Q3 2015

With the volatility of China’s stock markets, it is unsurprising that portfolio investment has fled. The exodus followed a record year for portfolio inflows, when investors rushed in to ride China’s bull market. But sentiment has clearly turned against the Chinese markets: in January, the quotas for China’s QFII programme decreased for the first time since the scheme was introduced in 2006.

The second-largest contributor to the upswing in outflows through Q3 was the “loans” component of the BOP, which accounted for 30 per cent of the change. It was also the largest overall contributor to outflows, at nearly US$200 billion. The majority of the negative balance in loans flows came from Chinese companies and banks paying off their dollar-denominated debt.

Dollar credit to Chinese corporations has soared since 2009, from less than US$250 billion to nearly US$1.2 trillion by the end of Q2/15, as many Chinese corporates arbitraged the difference between onshore and offshore rates and took advantage of the steadily appreciating RMB. This trend looks to have reversed decisively in Q3 of last year after the RMB’s depreciation in August and a steady narrowing of the interest rate differentials after six rate cuts in China from November 2014 to October 2015.

Because most of China’s dollar loans are offshore, we do not have detailed data; but in looking at the data we do have for onshore banks, we see clear evidence that Chinese corporates are scrambling to pay down dollar debt. Foreign currency loans in China, which had been on the rise for years, have turned sharply negative since August (see Chart 4 below), decreasing by close to RMB650 billion (US$99 billion) over the last five months of 2015.

Chart 4: Foreign currency loans by Chinese banks

Other signs suggest that Chinese companies have been paying down offshore dollar debt as well. The surge in China’s corporate bond market in 2015 is in part due to Chinese firms switching dollar debt into RMB debt.

Before 2015, most Chinese property developers were not allowed to issue bonds on the domestic market; in order to raise funds, they issued bonds abroad. After the government lifted restrictions on in January 2015, there was a surge in bond issuance, and part of these proceeds were then used to pay off foreign liabilities. The upturn in bond issuance starting in August is, we believe, in large part a response to a growing desire to pay off dollar debt as RMB depreciation expectations became entrenched after the 11 August devaluation (see Chart 5 below).

Chart 5: Onshore corporate bond issuance

That large amounts of China’s outflows are going to pay off dollar debt is further corroborated by a look at China’s external debt statistics. According to the State Administration of Foreign Exchange (SAFE), external debt decreased by US$150 billion during Q3/15, with almost all of that going to pay off short-term debt.

What all the above means is that most of China’s outflows actually signify a restructuring or refinancing of Chinese corporate debt, with many companies looking to borrow more in RMB and less in USD as interest rate differentials narrow and currency risk increases. This interpretation would seem to be further corroborated by China’s January 2016 lending statistics. They showed a dramatic increase in onshore loans to RMB2.51 trillion (US$390 billion), almost double the figure from a year earlier.

A sizeable minority (US$85 billion) of the “loans” outflows also comes from increased overseas lending by Chinese financial institutions, continuing a trend since 2010 of more Chinese lending abroad (see Chart 6 below). Its further acceleration in 2015 was probably due in part to Chinese institutions attempting to support Xi Jinping’s signature One Belt One Road initiative, which seeks to increase Chinese investment abroad. (For more, see our 24 November 2015 note China Global: What the ‘One Belt One Road’ initiative means.)

Chart 6: International loan assets of Chinese financial institutions

The third major driver of outflows to Q3 was “currency and deposits”, which in essence means that Chinese corporates and households are increasing their dollar holdings while foreigners are decreasing their yuan holdings. This is a very rational reaction to continued RMB depreciation expectations, which makes it a losing proposition for corporates to hold yuan. But it does not appear to be capital flight per se. Instead it is an indication of sound corporate management and a desire to rebalance assets. Likewise the increase in the negative balance of trade credits and advances reflects a desire to hold dollars and to spend RMB. As such, we do not see most of these flows as capital flight.

Finally it is worth noting that errors and omissions now account for a substantial portion of China’s outflows, much beyond what we would expect from simple statistical discrepancies. These look likely to have surpassed US$200 billion in 2015, and are clear evidence of Chinese corporates and households attempting to evade capital controls. Some of these flows are likely to be benign but the majority are likely to be capital flight.

As Chart 7 below shows, errors and omissions flows became more negative in 2012 in the run-up to China’s leadership transition. We believe that this was probably due to wealthy individuals looking to protect assets from seizure as a consequence of political factors. The desire to get assets out of the country has only increased in the wake of Xi Jinping’s anti-corruption campaign, which began in December 2012 and continues. (For more on the anti-corruption campaign, see our notes China Monitor: Xi goes after former leader's power base of 9 February 2016 and China Monitor: China’s anti-corruption probe reaches deeper into business – and is set to expand of 18 December 2015.)

Chart 7: Quarterly errors and omissions flows

Our examination of the BOP data to Q3 leads us to believe that at least two-thirds of China’s outflows through Q3/2015 can be explained by relatively benign factors, with relatively little representing capital flight as we traditionally understand it. In fact, most of the money is not leaving China so much as being shifted among entities within it. Most of the flows can be understood as a shifting of dollars off the central bank balance sheet and on to the balance sheets of Chinese companies.

However outflows in Q4 were potentially more worrying, as they appear to be driven more by households. In the run-up to the end of the year, financial advisors in China were urging their clients to use their US$50,000 quota for foreign exchange, and anecdotal evidence points to an upturn in households buying dollars. As dollar purchases increased, authorities told financial advisors not to encourage the practice, but it was too late. Dollar buying by households was strong again in January after the annual US$50,000 quota was renewed. Consequently, outflows in December and January were particularly large, totalling over US$300 billion.

This upturn in outflows from households is not necessarily bad, however, as we believe that most of these dollars will be converted back into yuan once depreciation expectations have ended.

Economic implications: It’s more about the currency than the economy

Whether it is households buying dollars or Chinese companies paying off debt, the motivating force behind the vast majority of these flows is expectations about the direction of the currency. What comes through very clearly in looking at China’s BOP data is that outflows pick up when the RMB is felt to be weak.

Net capital flows began to turn decisively negative in Q2/14 after the yuan started to depreciate earlier in the year. Outflows were still more or less offset by the positive current account balance until Q3/15, when the August devaluation spurred further outflows. Similarly, large outflows accelerated in December 2015 and January 2016 after the yuan started to weaken again.

Chart 8: USD/CNY and periods of increased outflows

However, negative expectations about the currency are not necessarily indicative of negative sentiment on the economy. FDI growth is slowing down but it is still large and positive. In the American Chamber of Commerce in China’s most recent business climate survey, over two-thirds of firms said they intend to increase their investment in China in 2016.

Animal spirits run high in China and, to some extent, the outflows are reminiscent of the way that money poured into the stock market in 2014 and the beginning of 2015. Just like the stock market then, the currency now looks like a one-way bet, and Chinese investors are flocking to take advantage. But once yuan expectations stabilize, we expect that the outflows will too.

The government’s response: Muddle through probably muddles through

Much like in their response to the rout in the stock market, Chinese authorities in late 2015 and early 2016 looked powerless to stop depreciation expectations and the outflows that are both causing and resulting from them. Things appear to have stabilized somewhat after People’s Bank of China (PBoC) Governor Zhou Xiaochuan explained the government’s thinking this past weekend, but there remains the possibility that depreciation pressures could resurface. (For more on Zhou’s recent comments, see our 17 February note China Reform Agenda Mark Two: Don’t run before you can walk)

The problem for policymakers is that it is difficult to stop depreciation expectations once they gain traction. If Chinese corporates and households continue to sell the RMB, the currency will be under further pressure to depreciate. The government can continue to maintain relative stability by using its FX reserves to buy back RMB; but this is hardly a solution, as it only reinforces the impression that the currency is artificially strong and thus adds to depreciation expectations.

Pressure to find a solution has mounted as supporting the currency has drained money from the domestic financial system at a time when liquidity is already tight. The PBoC’s ability to respond to liquidity pressures is constrained: in the context of a depreciating currency, it cannot cut interest rates because this would put further downward pressure on the currency.

For now the government has tried to put a brave face on the situation, and has expressed little concern over the outflows. Premier Li Keqiang said on 6 February that the currency would remain “basically stable”, and in January a SAFE spokesperson said that "the risks of cross-border capital flow are controllable on the whole." And on 14 February, PBoC Governor Zhou broke his five-month silence to give the most clear defence yet of the currency and the government’s strategy. He said that the underlying fundamentals driving the exchange rate (the current account and inflation) indicate that the currency is not severely overvalued. The government will continue to intervene to guard against speculation but the ultimate goal is “to have the exchange rate ‘broadly stable at an adaptive and equilibrium level’".

In the short term, therefore, we believe that the authorities are likely to exert stronger enforcement of capital controls while gradually depreciating the currency (and occasionally adjusting the fixing rate higher in order to hurt speculators). Once depreciation expectations have abated, we are likely to see a resumption of reforms to open the capital account and to liberalize (but not free) the exchange rate.

How big are the outflows?

Analysts disagree markedly about the size of China’s capital outflows. The two most often quoted estimates are those of Bloomberg and the International Institute of Finance (IIF), who have estimated the outflows in 2015 at US$1 trillion and US$676 billion respectively. We estimate that outflows were between these two estimates, at approximately US$870 billion.

These estimates diverge by so much for several reasons. First there is disagreement about what constitutes a capital outflow. Second we do not have all the necessary data to accurately assess the outflows. Third, for the data that we do have, there is disagreement about how to interpret it. We address below these three issues and explain the assumptions that underlie our estimate of US$870 billion.

Before we can address these issues, we need to construct China’s balance of payments data for 2015. Some data for Q4 have yet to be released; but because we do have data for Q1-Q3 and some of the important Q4 figures, we can create a picture of the annual data that is sufficient for our purposes. We know the two most important pieces of data for China’s current account: China recorded a goods trade surplus of US$578 billion and a services trade deficit of US$209 billion. Q4 data for primary and secondary income are unavailable; but based on data for earlier in the year and China’s historical figures for this data, we estimate that it was somewhere between -US$67 and -US$107 billion. This gives us a current account surplus of US$262-302 billion. For our purposes, we use a median figure of US$282 billion.

Table 2: China’s 2015 balance of payments (US$ billion)

As we also know that China’s FX reserves decreased by US$513 billon, the capital and financial account (which we will from now on refer to simply as the financial account) and errors and omissions must total US$795 billion. Assuming that all errors and omissions account for capital outflows (a reasonable assumption in this case), this would give total net capital outflows of around US$795 billion. We have assumed that of this amount, approximately US$595 billion came through the financial account, with the remaining US$200 billion coming through errors and omissions. However we caution that these figures are preliminary, as the relevant Q4 data are yet to be released.

But this figure of US$795 billion is only a starting point. Because a portion of China’s reserves are held in euros and sterling, both of which depreciated against the dollar in 2015, we need to factor in valuation effects for the overall decline in reserves. The composition of China’s FX reserves is undisclosed – it is regarded as a state secret – but we estimate it to follow roughly the IMF’s Currency Composition of Foreign Exchange Reserves ratios of 67 per cent USD, 20 per cent EUR, 5 per cent GBP and 5 per cent JPY. If this is the case, then roughly US$100 billion in reserve losses would be attributable to valuation effects. This means that we should adjust our estimate of outflows downward to US$695 billion. We believe that this is more or less how the IIF arrived at its estimate of US$676 billion.

The above is a well-accepted method for calculating outflows but there are other approaches. Many, including Bloomberg, do not count direct investment as a capital flow. Because China runs a sizeable direct investment surplus of around US$100 billion – we estimate a surplus of US$88-128 billion – subtracting it from our capital flows calculation would increase net outflows to US$795 billion.

We do not agree with those who do not include direct investment in calculations of capital flows. In our view, capital used for investment represents a capital flow, whether that money is used to acquire a company (direct investment) or to buy shares in a company (portfolio investment). However there is another, more reasonable argument to be made for disregarding China’s direct investment figures in calculations of capital flows. China’s direct investment surplus is almost entirely the consequence of the difference in reinvested funds between foreign companies in China and Chinese companies operating abroad. This is due to the much higher stock of foreign investment in China, where foreign companies have been investing heavily for decades, as compared to the stock of investment by Chinese enterprises outside China, which have only recently begun to increase their international footprint on a large scale. We think it reasonable to exclude these reinvested funds from China’s capital flows, as they do not represent money coming into China in the way we typically think of capital inflows.

So to calculate the flow of funds related to direct investment, we need to look at the money actually flowing into and out of China. Luckily the Ministry of Commerce (MoFCOM) calculates FDI and ODI in this manner, and its figures show there to be a slight direct investment surplus, with FDI in 2015 of US$126 billion and ODI of US$118 billion. For simplicity’s sake, we will call this a wash as US$8 billion is insubstantial in the context of China’s capital flows. This means that our calculations are line with those of Bloomberg and others who ignore direct investment, even though our methodology is different.

One reason that Bloomberg and others tend to have much higher estimates of capital outflows is that they fail to consider the effects of China’s services trade deficit in their calculations. We do not know why they do this, but it seems a particularly egregious oversight in that China’s services trade constitutes an ever-larger portion of China’s current account balance. The services trade deficit grew to -US$209 billion in 2015 from -US$151 billion in 2014 and -US$124 billion 2013 (see Chart 9 below). By ignoring the net US$200 billion spent on imported services, these purchases get counted as capital outflows, which is how Bloomberg and others get to total outflows of US$1 trillion.

Chart 9: Annual services trade balance

While we do not think that the services trade should be completely discounted, we do believe that it should not be taken at face value. It includes “travel” imports, which reflect any time that Chinese use credit cards or withdraw money abroad; these doubtless conceal some outflows. It is also widely known that overinvoicing for imports is one of the most common ways to get around China’s capital controls and that increasingly it has been concentrated in the services trade. This has long been a problem in China, but the scale seems to be much bigger this year. The government was so concerned about this practice that it issued rules against it in September once China’s outflows started to accelerate markedly.

We estimate that US$75 billion of outflows might have been miscategorized as trade flows in the official data, bringing our house estimate for capital outflows in 2015 to US$870 billion. We also point out that if some of China’s capital outflows were miscategorized as imports, then China’s trade surplus was actually bigger than the reported figure, so these flows would not affect China’s overall net FX position.

A few final caveats on calculating outflows: distortions and/or statistical errors in the data are inevitable when trying to track flows of this magnitude – all the flows recorded in the BOP total well over US$6 trillion. Exchange rate fluctuations and other distortions could swing the total outflows count by tens of billions of dollars in either direction.

There are also non-negligible discrepancies between the trade data recorded by MoFCOM and that recorded by SAFE. MoFCOM reports a merchandise trade surplus of US$595 billion and a services trade deficit of only $US137 billion; SAFE’s numbers are $US578 billion and $US209 billion respectively. We have used the SAFE figures in our calculations; but had we used the MoFCOM figures, this would imply further outflows of nearly US$90 billion.

Finally all this data will undergo further revision. As such, the margin of error for all these calculations is quite high, and could easily be US$150 billion in either direction.

What lies ahead: The edge of the woods is near

We believe that China’s capital outflows are likely to slow down significantly in the coming months. This is because the household outflows that drove outflows in December and January will decrease as annual FX quotas are used up and the authorities step up the enforcement of capital controls. We also believe that much of the previously unhedged dollar debt has now been paid down or hedged and that Chinese corporates have already shifted as much cash as possible out of RMB. Most important, expectations of further dollar strengthening seem to be reversing, meaning that depreciation pressures could quickly abate.

While our base-case scenario sees outflows and the RMB stabilizing, there is still a chance that things could turn ugly. Currency crises, by their nature, are sudden and unexpected (even if they look inevitable in hindsight). The key indicator to watch will be the official foreign reserves. Estimates vary on how large reserves need to be in order to support the currency. They range from US$1.5 trillion at the low end to US$3 trillion at the high end. There is also debate about how much of China’s reserves are actually available to defend the currency.

What matters is what Chinese households and, to a lesser extent, corporates think. If they become convinced that the currency is grossly overvalued and that the government is powerless to defend it, they will attempt to sell RMB en masse and a currency crisis will ensue. If net outflows continue to run at the US$150-200 billion/month rate that we have seen recently, this could happen within the year. For now, this seems a remote possibility, but just one year ago it did not seem a possibility at all.

Recent TS Research