Brad Setser

Follow the Money

Cross border flows, with a bit of macroeconomics

Print Print Cite Cite
Style: MLA APA Chicago Close


Do China’s fast growing dollar reserves guarantee a sound banking system?

by Brad Setser
May 26, 2006

One common argument about China that I never have fully understood is that China’s banks are OK because China has tons of dollar reserves.

It is implicit in this statement by the (very good) Richard McGregor in his FT report on the (retracted) Ernst and Young report on China’s bad loans.

China's total liabilities for non-performing loans may be as high as $900bn, dwarfing official estimates and outstripping the country's massive foreign exchange reserves, according to a study of Beijing's bad debt problem.

I don’t get it.  

Foreign exchange reserves are useful if depositors want to take their money out of the country.   But when it somes to making up a gap between Chinese banks RMB deposits and their RMB deposits, they need an RMB asset – not dollars.    Typically that asset is a government bond.  So what matters – far more than China’s dollar reserves –is the capacity of China’s government to issue and pay a bunch of RMB debt. 

I am assuming, of course, that China’s banks are not totally sound now.  No doubt, a lot of bad loans have been bought by the PBoC and shifted to the asset management companies (which will need to be bailed out) or otherwise moved off the banks' books to prepare three of the big four state commercial banks for stock market listings.  But a certain fraction of their new loans are also likely to go bad.

But the quality of the banks balance sheets is a seperate issue from how you bailout bad banks. 

When the US bailed out its savings and loans, it didn’t use its euro and yen reserves.

When Sweden bailed out its banks, it didn’t use its dollar reserves.

When Japan bailed out its banks, it added to its dollar reserves.

When Indonesia (and others in Asia) bailed out their banks, they did so in local currency.  Indonesia had no choice: it lacked sufficient dollar reserves.

If the banks deposits are in the local currency – and China’s are – the banks need a local currency asset, not foreign currency reserves.

Hell, when Argentinabailed out its banks – which held a lot of dollar deposits — it first converted dollar deposits into peso deposits and then issued peso denominated bonds to cover the banks’ losses (this is a vast over simplification).  If the banks needed pesos, not peso-denominated bonds, they could get those from the central bank as well.  Argentina had no real choice: it didn’t have enough dollars left at the end of 2001 to back the banks dollar deposits.

Bank recapitalization is something that lots of folks don’t understand, but it is really sort of simple.

Say a bank takes in say RMB deposits.  It makes RMB loans.  Some loans go bad.  The bank has more RMB deposits than good RMB loans.  If those bad loans exceed its capital, it is technically bankrupt.  But if the bank is owned by government, the government usually will make up the difference by giving the banks a RMB bond. In a crisis, the government generally will do the same thing for a private bank.  Sometimes the government takes control of the bank as well.  Sometimes it doesn’t.

The RMB bond assures that the banks assets (a mix of bonds and performing loans) are equal to its liabilities (RMB deposits).  And since both the bank’s deposits and its assets are in RMB, the bank is matched, currency-wise.

That is the problem with giving the banks the country’s foreign exchange reserves.   It creates a currency mismatch on the banks’ balance sheet.  Potentially a big one.   RMB deposits need to be matched with RMB bonds and RMB loans.  Not dollars or Euros.   If the banks get dollars or euros, and the dollar or euro depreciates against the RMB, good banks will become bad banks quickly.  Depreciating assets are not a good thing for a bank.

I think part of the confusion stems from the fact that China has shifted $60b of reserves to three banks to help them meet their international capital adequacy standards.  That capital doesn’t directly back deposits.   And it isn’t perfect from the banks point of view, since they are left with the exchange rate risk — though the central bank reportedly has promised it will protect the RMB value of their dollar capital. 

But it is important to remember that this transfer of reserves is only a small part of China’s bank recapitalization.   In general, when bad loans have been moved to the AMCs (or bought by the PBoC) the banks have been given RMB, whether cash or an AMC bond – not dollars or euros.   That is as it should be.

There is one set of circumstances where China’s bad banks could require China to use its dollar reserves.  It goes like this.

Chinese bank depositors lose confidence in China’s banks, and start to withdraw deposits from the banking system in mass.  That would be a big change from the very strong deposit growth we are observing right now.   But it could happen.

The banks would first draw on their stock of liquid RMB assets.  And if that wasn’t enough, they could sell their RMB recap bonds to the central bank for RMB cash.

As depositors pulled their funds out of the banks, they would end up holding a huge stash of RMB cash.   And they might want to convert that to dollars or euros.

China has capital controls, so this isn’t easy.  But let’s suppose for the sake of argument that the controls are lifted.   The central bank maintains a de facto peg – so the depositors could sell their RMB to the central bank for its dollars. 

Bingo –

Two key points:

First, the dollars are useful if Chinese citizens want to pull their funds out of China.  That is not quite the same thing as bank recapitalization — even though concerns about the quality of the banks are one reason why Chinese citizens might want to pull their funds out.  That said, Chinese citizens had far more reason to pull their funds out of the banks in 2002 than they do now.   A lot of dud loans have been shifted to the AMCs over the past two years (particularly from ICBC).

Second, with a current account surplus of $150b (and growing) and net FDI inflows of $50b, Chinese deposits could send up to $200b a year without requiring the central bank to dip into its existing reserves at all.   Rather than financing reserve buildup, China’s enormous surplus in its basic balance of payments would just finance capital flight.

The fact that China needs a capital outflow of nearly 10% of its GDP – the kind of outflow that we saw in Argentina in the peak of its 2001 crisis — just to keep its reserves from growing is one reason why I think the argument that “we don’t know what would happen to the RMB if China lifted its capital controls because its banks are so bad” is a bit overstated.  

We basically know that China needs an Argentine crisis level of outflow from the banks just to keep its exchange rate from appreciating, given the strong pressure for appreciation stemming from its trade surplus and FDI inflows (I'll deal with Stephen Green's argument that the Chinese data overstates China's current account surplus and understates capital flows into China at another time — but one key point is that folks want to get into China, not get out).  Actually, China needs more than just a one-off Argentine style crisis.  It needs an ongoing Argentine-style crisis that generates a 10% of GDP capital outflow every year.  If a big capital outflow just happens in one year that isn’t enough, since the next year, Chinese exports and FDI inflows will bring in another $200b plus …

I don’t think that is likely – and I am not among those who think China’s banks are in all that good a shape.  But that is another topic.

As is the possibility that rapid reserve growth — by leading to rapid money and credit growth and by forcing the government to rely on administrative controls rather than markets to limit lending could actually be hurting the banks long-term health.   Even if the current lending boom — particularly with a government mandated ceiling on deposit rates and floor on lending rates — is dramatically increasing the banks' current profitability.


  • Posted by Dave Chiang

    Chinese per capita Debt levels low in comparison to Europe and the US

    Statistics released by the Budgetary Work Committee of the NPC Standing Committee showed that by the end of 2004, China’s national debt balance reached 2.96 trillion yuan (US$370 billion), including 2.88 trillion yuan of domestic debt and 82.8 billion yuan of foreign debt. China’s 2004 national debt burden rate was 21.6% of gross domestic product (GDP), far lower than the warning level of 60% designated by the European Union for its members, a generally accepted international standard.

    By comparison, the US national debt stood at $8.4 trillion as of April 13, 2006, or 65% of forecast GDP. About $4.9 trillion of the US national debt is held by the public and $3.5 trillion is held intra-governmentally. US national debt is about 20 times China’s on a nominal basis, five times on a purchasing-power-parity basis, almost four times on a debt-to-GDP basis, and 100 times on a per capita basis. US per capita income is about 35 times that of China in 2005, which means each US citizen is carrying almost three times the national debt-to-income ratio as his or her Chinese counterpart.

    – Asia Times

  • Posted by Guest

    I have merely a question that someone might be able to answer. The Bank of China has a Hong Kong branch or subsidiary (Bank of China Hong Kong) that would appear to be sound and well managed. Anyone know if it has bad loans of the sort that exist in mainland banks? And another Hong Kong bank, the Bank of East Asia, would appear very well capitalized with equity around 16% of assets (and a rather low return on equity as a result). Is it also burdened with bad loans, or are these Hong Kong based banks in far better shape than mainland Chinese banks?

  • Posted by Joseph Wang

    Chiang: Apples and oranges. That figure doesn’t include the $200 billion or so of bad debt that are in the AMC’s which will have to be funded from the taxpayer, and I’m guessing that there are about $50-$100 billion in the rural cooperatives that will also need a bailout eventually. There is also probably about $50-$100 billion in hidden municipal debt.

    If you add those items, you get a debt burden which is still reasonable (40-50% GDP), but you need to get the accounting right. I’d argue that if the money went into things like improving social infrastructure that debt isn’t a bad thing.

  • Posted by Joseph Wang

    Guest: The problem with the big four banks was that they ended up being social welfare funding agencies to provide unemployment benefits to ease China’s transition to capitalism. (Which is why I think that the debt associated with it should be counted as “public debt.”)

    Personally, I don’t think this was a bad thing. In hindsight, it may have been the least bad of the alternatives. Lardy has argued that China would have been better off funding social welfare benefits of the transition straight from government revenues, but the nice thing about funding them through NPL’s is that you couldn’t do it forever.

    The HK branches of the BOC never had this problem.

  • Posted by Charlie

    As long as bank deposits keep growing at a good clip, bad loans can be hidden. The only way to tell the true health of their banking system is after a period of time where bad loans are growing at a faster rate than net deposits. As long as China’s economy continues to expand at a high rate, the bad loans will likely remain hidden.

    Common sense would dictate there are probably a lot of bad loans in China. There has got to be a lot of overcapacity in certain sectors (building materials comes to mind) along with mismanagement. Whether the total is closer $100Billion or $1Trillion is debatable.

    If a banking crises were to occur, my bet is the chinese government would loan Yuan to the banks.

  • Posted by Joseph Wang

    Charlie: The trouble is that in the mid-1990’s, the losses from state-owned enterprises that the banks were being asked to bail out were increasing. The total amount of bad loans in 1998 was manageable, but the trouble was that without some major economic changes, the rate at which those loans were increasing was not.

    Also, the fall of the Suharto government in Indonesia shocked the Chinese government into action as far as fixing the banks. There *will* be some sort of financial crisis at some point (just like there *will* be a major hurricane hitting Florida at some point or an earthquake hitting California). Much of the goal of Chinese banking reform is to make sure that *when* the crisis hits, that it won’t take down the economy and the Communist Party with it.

    It makes a big difference whether there are a $100 billion in bad loans or $1 trillion in bad loans. The former is an annoyance. The latter is a huge pending crisis. Part of the reason I’m optimistic is that if you look at the sectors that are having overcapacity problems (real estate), you really don’t see that many bank loans being allocated there.

  • Posted by HZ

    Bad loans’ effect on money supply is indirect. Of itself, bad loans merely transfer ownership of money. As time goes by, the reduced capital stock limits banks’ ability to lend, and so far as banks are the most leveraged entities, decreases the leverage in the credit markets. If a run on the banks develops (though it is inconceivable such things happening to state banks), good loans may have to be called in and further depress leverage. Recapitalization is to counter such deflationary pressure. Financially it makes little difference how state banks are capitalized. The moral hazard is always there: the nation as a whole bear the cost of recapitalization, while only a few benefited from the bad loans directly. The whole issue is a distribution/allocation problem.
    That said, Chinese banks are very much under-leveraged, so it is hard to see how bad loans can come to a head, which probably also explain their propensity to further accumulate bad loans.

    I find Stephen Green’s argument very interesting. I’ve made similar points before but I didn’t have any numbers. Stephen stuck his neck out by putting numbers on the table. I would be very interested in Brad’s comments on the issue.

  • Posted by Lord

    How about if they adopt the dollar and retire the RMB?

  • Posted by pgl

    But couldn’t the Chinese banks convert $ to yuan? Oh yea, that might trigger a chain of events that would force currency adjustments. But on a simpler topic – does the fact that E&Y can’t even audit Chinese banks cause one to wonder about the accounting profession?

  • Posted by Guest
  • Posted by Joseph Wang

    The Ernst and Young problem happened because you have accounting firms doing too many things. The report itself on Chinese NPL’s was an “infomercial” issued by the consulting branch of E&Y, and they obviously didn’t talk to the accountants before issuing it.

  • Posted by OldVet

    Joseph, “they obviously didn’t talk to the accountants before issuing it.”

    That was also the problem with Arthur Andersen and Enron, wasn’t it? Failure to put distance between E&Y and client is similar.

  • Posted by Anonymous

    Let’s get real … we are talking about nearly $1 trillion in bad loans … E&Y is right in calling out that this is going to make a material splash … they are wrong to simply connect it to reserves … the only reason to do so is for rhetorical purposes of demonstrating magnitude … the magnitude comparison is reaonable given that China is not playing with a small game on either bet …With that said, Brad, the number is HUGE and you should take it seriously … all your hand waving is silly … Joseph, you are normally a light of reason … thi is a HUGE number

    This problem is too LARGE not to make a splash … Brad you don’t have to go on barking about Argentina … Asia had a spell with bad loans … it created a MESS … I will ask you also to think Japan to see how bad loans can ravage an economy for decades. Brad please you a little more reasoning … stop with teh Argentina story … its wering thin and Asia has its own stories

  • Posted by Joseph Wang

    It’s easy to get lost in large numbers but one trick is to use megabucks, gigabucks, and terabucks.

    A 100 gigabuck problem is very different from a 1000 gigabuck problem.

  • Posted by Joseph Wang

    To give you a sense of scale, Walmart’s 2005 total revenue was about 280 gigabucks. Citigroup had 100 gigabucks in revenue off assets of 1500 gigabucks.

    Which goes to my point that if the Chinese NPL problem is a 100-200 gigabuck problem, then it is annoying but not a crisis. If it is a 1000 gigabuck problem, then it is an impending crisis.

    The other thing to keep in mind is that the goal of the Chinese government is obviously not merely to create banks that are “not bankrupt” but rather ones that are “world class.”

    The long term strategy of the Chinese government is to modernize the big four banks to the point where CCB, ICBC, BOC, and even ABC are mentioned in the same breath as Citigroup or JP Morgan-Chase.

    I have heard someone speculate that the real reason that Chinese banks are inviting foreign investment isn’t for the money or the expertise, but rather so that Chinese banks have some very high power political allies in about ten years when they are ready to set up operations in the United States.

    If you thought sparks were flying in the CNOOC battle, just wait a bit……

  • Posted by Guest


    The fact that the PBoC has $900 billion USDs in reserves makes a huge difference. If the PBoC had $20 billion in reserves, then the PBoC would have to issue bonds to cover this problem with no backing whatsoever –a la Argentina with a huge premium…

    Central Banks are the ones that come to the rescue of banks when they’re in trouble, basically by lending them money –CNYs, USDs, whatever is required. In exchange they get a promisory note (bond) for the terms of the repayment.

    Bottom line, it’s better to have reserves while the PBoC is doing this than not.
    Argentina comes to mind — 3+ times currency devaluation, huge interest rates to capture USDs for their CB’s bonds…

    With $900 billion USDs to back this up, the PBoC can issue (good quality, lower rate) CNY bonds and get (poor quality, higher rate) CNY bonds from the affected banks; whom in exchange receive the CNYs to cover their NPLs.

  • Posted by HK

    Brad–I completely agree with your analysis; foreign exchange reserves do not help solving NPL problems in the banking sector, although they would help avoiding a currency crisis like Indonesian or Argentine one (which is extremely unlikely anyway). Chinese NPL problems can only be solved by substantial capital injection by the government and strengthened bank regulation and supervision.

    The injection of $60 billion foreign exchanges does not make much sense, as you described aptly. However, this was a compromise reached between the Ministry of Finance, which did not like budget money used outright for bailing out banks, and the People’s Bank of China, which was seriously concerned about NPL problems in large state-owned commercial banks. Eventually, those $60 billion would be lost. Who bears the immediate loss, MoF or PBoC, seems to be an unsettled political issue, but financially does not matter much since the government (eventually the people) bears the loss anyway.

  • Posted by LWong

    BSetser and JWang – The Central Government(and hence the People’s Bank of China) has a huge asset if it ever needs to raise terabucks of capital – landownership rights. At the moment, only land-use rights have been sold to individuals/corporations.

  • Posted by DF

    Dave Chiang, public debt is a tiny fraction of the overall problem. Bad loans are mostly loans to private agents, households and companies.

    The problem is not the US federal debt of around 60% of GDP it is the US total debt of around 250% of GDP, with household debt at more than 100% and rising.
    It’s the same problem in China.

    When all these potential non performing loans are taken into consideration it becomes self evident that only one result can be expected : Debt deflation.

    There’s no way any central bank can work it’s way out of deflation if over a 10 year period the debt/GDP ratio is to fall 100%.

    Back to China, given that its export growth has to slow and its banking system is unsound … I don’t see much future in their growth … As soon as the US consumer catches a cold I think the Chinese worker is going to catch the asian flu.

  • Posted by Joe Rotger

    (I’m the previous guest…)

    I further wanted to stress that having substantial reserves allows a country to have a substantial level of independence in handling their internal affairs, rather than be controlled by the overbearing requirement to have hard currency to trade with the world.

    In essence, if a country has small CB reserves, an unexpected high level of NPLs can rapidly escalate to a desperate situation and has a good chance of becoming catastrophic.

    I subscribe to the notion that passing along bonds to the ailing banks will certainly make the NPLs persist in the future, unless these banks policies are reformed.

    Although, I have the feeling that there’s a lot more involved –it would probably require the government to lose significant control on lending… Can this be achieved without a major political change? It looks like a very steep slope to climb…

  • Posted by bsetser

    Guest — actually, Argentina has generally issued CB bonds at relatively low rates after its devaluation, in part b/c it hasn’t sterilized much and in part b/c the banks are relatively liquid.

    I agree that fx reserves are useful if (and tis a big if) there is a run out of banks into foreign currency.

    I disagree with the idea that fx reserves are useful when it comes to recapitalizing the banks. HK is right. And right on the MoF/ PBoC policies. MoF’s reluctance to use fiscal resources to recap the banks is leading to all sorts of (in my view) mistakes, including overuse of AMCs.

    Steve W. Interesting thought. Tho right now the PBoC is effectively monetizing both bad loans (by buying NPLs from ICBC for cash at par) and fx reserves. I am not sure though that the net effect is positive — lots of domestic liquidity that either had to be looked up in the banks or not.

    As for it is better to have reserves than not, two points:

    a) I take the PBoC’s concerns about the difficulties mopping up all the liquidity created by all of the reserve growth seriously. Too much reserve growth = too much money = too many deposits = too much lending chasing too few assets … and so on. The process of adding to your reserves indirectly can help generate new bad loans, since reserves are purchased with cash and thus in the first instance lead to more money growth (and the PBoC doesn’t fully sterilize)

    b) relative to their total loans, Chinese banks had more NPLs and China had fewer reserves in 2002 … I consequently am not convinced that China needs $900b plus in reserves to guarnantee confidence in its banking system through the indirect measures laid out above (i.e. so long as depositors know thay the reserves are there so that they could shift from RMB to $, they are happy not just to hold their existing RMB, but add to their RMB balances … ).

    Anonymous with a nasty tone (let’s get real, stop using argentina … by the way, I am less than impressed by those who use anonymity for critical comments; i happily accept criticism, but prefer it some from someone who can be identified at least by a internet handle for future reference). I never have denied that China has an NPL problem. I am only arguing that reserves do not directly help solve that problem. Japan is a good case in point. it had tons of reserves and its banks had tons of bad property loans. Did it ever use those reserves to buy the banks bad property loans to help clean up the banks? No — that would have created a big currency mismatch for the banks. Rather than using its reserves, Japan added to its reserves (supporting its export sector) during the period when it was cleaning up the banks.

    Like I said, reserves are useful if there is a run out of the banks into foreign currency — i.e. problems on the liability side of a bank balance sheet. they aren’t of much help cleaning up the asset side of bank balance sheets. And the NPL problem is on the asset side. the MoF needs to write a big RMB check.

  • Posted by Gcs

    no time to read coments though i see some legendary long ball hitters in the line up…
    but two quick points

    sorry if i repaet others


    are you forgetting the full implication
    of “the prc can make all the rmb it needs ”

    what makes policy max
    allocation of credit inferior to profit max allocations

    in a real world second best market system
    where prices can give off wrong signals
    like the devil at an intersection

  • Posted by Anonymous

    What is the most interesting reading this blog is to see for how long Brand and Joseph have sounded intelligent but provided no material insight into the game at hand. They yap a lot about the game, but never call a single game correctly. They sound like sport announcers or the talking heads on cable.

  • Posted by Joseph Wang

    Anonymous: If you want a specific prediction ask for one. The big prediction that I’ve made over the last few years is that the Chinese economy isn’t going to collapse. (I also predicted before the war, that there would be no WMD’s found in Iraq.)

    LWong: I doubt that the PRC central government can raise very much money from land sales. Urban land has already effectively been sold off, and if you try to sell off rural land to raise money, you are looking at total revolution. The other thing is that money from land sales go to local government coffers and not the central government.

  • Posted by MTC

    Dear “Anonymous” –

    Perhaps you misunderstand. The reason Dr. Setser may not be providing material insight into the game “at hand” (thank you for making this qualification) is that the game is onanistically playing itself. When you ask Dr. Setser to “get real” you are asking him to do quite the opposite–you are asking him to join others in self-delusion. The pressure to join is strong–the powers within Morgan Stanley have finally broken Stephen Roach–but cannot alter the conclusion Dr. Setser draws from an honest assessment of the flood of hard data he provides to all for free.

    If you want to know about the game “at hand”, visit PC’s blog. He/She seems to have a strong sense about where the market is heading.

    Mr. Wang rarely offers analysis one can translate into immediate economic action. Blaming him for not providing material insight into the game at hand is unfair.

    Oh yes, one more thing: you spatter the comment field with insults but choose to hide your own identity. What are you afraid of? You can trust our little comment community with your precious contact information. All here are fairly reasonable individuals, Dave C. excepted, of course.

  • Posted by DOR


    Two major changes in East Asia since 1997-98: larger forex reserves and China. The reserves are a critical cushion, but China is the real factor. China takes all the heat, takes all the speculative hot money, and that gives the other economies in the region a serious breathing space.

    Net result? My instinct is that if all of China’s money controls – capital movement, exchange rate, deposit and lending rates — were lifted, the net result would be a flight of money from the banks, but not from the country. One reason is faith in the future; the other reason is that the individual’s capital is so tiny that it would get eaten up in transaction fees in a second.


    BOC and BEA are both listed on the Hong Kong stock exchange. Both are in far better shape than anything in the Mainland of China. But, I don’t know about their debt levels (and, like all banks, I’ll bet they don’t know, either).


    A 747 is huge, but it isn’t unmanageable. The key is manageability.

    Joseph, Dave,

    The rule of thumb that held throughout the Asian Financial Crisis is that the first NPL-to-GDP figure posted is 1/3 of the final figure. If I remember correctly, China’s first figure was 14%.

    Gigabucks . . . I love it!


    No one worries about the accounting standards in China. Everyone knows they are useless, so why worry? P&L is pure PR.


    Nice explanation of how the world works. Very nice!