Brad Setser

Follow the Money

Cross border flows, with a bit of macroeconomics

Print Print Cite Cite
Style: MLA APA Chicago Close


Why China’s $1 trillion in reserves are unlikely to be of much use in a banking crisis

by Brad Setser
September 21, 2006

It is rather hard to read Andrew Browne’s report of the building boom in Zhengzhou in last week’s Wall Street Journal — or for that matter many other accounts of China’s current investment boom – and not come away thinking that the banks financing various mega-projects in various Chinese provinces won’t end up with a lot of non-performing loans.  Not every interior city will be the “Chinese Chicago” — let alone the next Shanghai.

I do think the improvement in the balance sheets of the Chinese banking system over the past few years is real.   Tons of old bad loans (ones made to SOEs in the 1990s) have been moved off the banks balance sheets.  Read Guonan Ma.   Or read my article – but my data came from Ma. 

Those bad loans are now parked in China’s Asset Management Companies, often after making a brief stop-over on the balance sheet of the People’s Bank (the PBoC has bought its share of bad loans at par).   

Most of the loans that the banks have made during the current lending boom are performing.   So relative to where they were say five years ago, the banks are in far better shape –

The risk is that a lot of these new loans will only perform so long as China’s boom continues.  And booms usually don’t last forever. 

Most people realize this.  They worry about the future health of China’s banks.

And a lot of people think this fact is a good reason for China to be holding so many reserves.     Friday’s FT leader, for example, suggests that China’s huge reserves might come in handy if a lot of the loans now being made turn bad. 

“Some China watchers believe the sector's bad loans are so vast that the government may need every cent of its trillion-dollar reserves.”

I disagree. External assets — like reserves — aren't what China will need in a domestic banking crisis.  Let me see if I can explain why in a convincing way. 

I suspect that China’s rapid reserve growth is adding to the risk of a banking crisis in China, not reducing it.   Rapid reserves growth is fueling a huge expansion of China’s money supply, as the PBoC hasn’t been able, or willing, to fully sterilize China’s reserve growth.  The result: a huge surge in bank deposits – and in the banking system’s ability to make loans.   The PBoC though doesn’t want the banking system to finance every provincial governors’ big dreams.  For that matter, the PBoC doesn’t really want the banks to lend out all their deposits.    But the banks make far more money (at least in the short-run) if they lend than if they do not.  So the PBoC is constantly struggling to keep the banks – which are flush with cash — from lending too much.   And it often fails.

Moreover, holding the exchange rate down means holding Chinese interest rates down – and directing Chinese monetary policy toward maintaining exchange rate stability, not toward curbing the domestic boom bust cycle.  That too has contributed to the current lending boom.

I am not the only one who thinks this either.  Far more impressive vocies have made a similar point.  Mike Mussa for one

The People’s Bank of China (the Chinese central bank) finds it difficult to sterilize all of the monetary effect of the massive increase in its foreign exchange reserves, thereby contributing to rapid money and credit growth.   This in turns fuels massive investment spending (as firms rather than consumers are the main recipients of bank credit.”

The Chinese leadership supposedly agrees (even if they haven’t been willing to take much real action). Washington Post, drawing on an interview with He Fan (of the Chinese Academy of Social Sciences).

China's leaders view booming exports and the resulting pile of foreign exchange reserves, now nearly $1 trillion, as added fuel for unwanted bank lending that is exacerbating the pace of investment. China's leaders have come to see a faster appreciation of the currency as a useful tool in the effort to slow investment and stave off trouble.

More importantly, I don’t think the dollars and euros that the PBoC is adding to its portfolio also will be of much use should boom turn to bust.    China isn't building up something in good times that will help it whether the bad times associated with a domestic banking crisis.  China’s government will not finance future bank recapitalization by drawing down its external reserves.  Rather, it will finance future bank recapitalization by issuing new domestic debts (effectively substituting a government bond for a bad loan on the bank’s balance sheet).


Reserves are an external asset.  External assets can be used to make up for a shortage of exports relative to imports.   A country with tons of reserves, for example, could finance a current account deficit by running down its reserves rather than by borrowing from abroad.     China doesn’t exactly have that problem.

Reserves are also useful if more capital wants to leave a country than come in.   That is why reserves can come in handy in a banking crisis. 

Not because most emerging markets use their reserves to recapitalize their banks after the banks make a series of bad loans.  No one generally does that. 

But if bank depositors want to withdraw funds from the domestic banking system and move their funds abroad, the country can finance that capital outflow by running down its reserves. 

The precise mechanism depends on the nature of the country’s banking system.

If the bank deposits are in dollars, the central bank can lend out its reserves to the banks – acting as a lender of last resort in hard currency.   The central bank provides the banks with the hard currency cash the banks need to pay their depositors, who wants dollars cash, not a dollar-denominated bank deposit.  The balance of payments accounting on this is hard (though working on Argentina in 2000-01 provided a good crash course).  A dollar-denominated bank deposit in the banking system of an emerging economy is a domestic asset while a dollar in cash is an external asset … so shifting from dollar denominated deposits to dollars cash generates a capital outflow.  Think of it in the following way: citizens are substituting a claim on the US (a dollar bill) for a claim on the local economy denominated in dollars (a dollar deposit that is used to finance a local dollar-denominated loan). 

All this doesn’t apply to China though.  Its domestic bank deposits are denominated in RMB, not dollars. 

If bank deposits are in local currency, the central bank doesn’t need dollars if its citizens want to withdraw funds from the banking system and hold cash – local currency cash – instead.  The PBoC can print all the RMB it will ever need. 

China would only need to draw on its dollar reserves if (and this is a big if) China’s citizens did not want to hold RMB cash — or to shift their RMB from a risky local bank to a safer local bank — but rather wanted to hold dollars (cash) or dollar-denominated offshore bank accounts.   

In theory, Chinese depositors don’t have the option of moving funds offshore.  China has capital controls and all.  But China’s capital controls leak.   We know that in 98 and 99 hot money flowed out of China.  If depositors lost confidence in the banks and in the RMB, those outflows might resume. 

But that doesn’t necessarily imply China would need to dip into its reserves. 

Why not?   Remember, China has an enormous trade and current account surplus.  Its 2006 surplus is estimated at $220b by the World Bank.  The August data suggests that the World Bank estimate may be low. Nick Lardy is now estimating a current account surplus of 9% of China’s GDP.    China would only need to dip into its stockpile of reserves if the total outflows exceeded the inflow from the current account surplus.   Actually, it would only need to draw on its reserves if the outflow from the banks exceeded the combined inflow from the current account and FDI flows.

Most of the emerging economies that had to dip heavily into their reserves when they had a banking and currency crises were initially running current account deficit, not surpluses.   They needed to draw on their reserves while their current account adjusted.   

Moreover, a banking crisis would likely be associated with a slump in new lending and slump in investment.   That that would tend to push China’s current account surplus up.  

The problem with running a 9% of GDP current account surplus during an investment boom is that it implies a far larger surplus in an investment bust.    China’s current account surplus would likely rise to around $300b in Chinese investment slowed, reducing demand for imports. 

Consequently, I would bet that China’s reserves would keep on rising even if China’s investment boom turns to an investment bust.  China would cling to its peg even more tenaciously to support its exports as the domestic economy slowed.   Barring enormous capital outflows, that the PBoC would continue to buy dollars to keep the RMB from appreciating …

Couldn’t China’s existing reserves be used to recapitalize China’s domestic banks? 

The answer is not really.

Why not?  

Simple.  China’s banks take in RMB deposits and need to match those deposits with assets denominated in RMB.    Right now, those deposits are matched with RMB loans.  It those loans go bad, the banks will need another performing RMB asset.

Dollars and euros, for all their respective virtues, are not RMB. 

Giving the banks dollars or euros would create a currency mismatch in the banking system.   That is a problem.  Particularly since dollars and euros are likely to fall in value relative to the RMB over time. 

The need to avoid a currency mismatch is a key reason why the vast majority of China’s recent bank recapitalization has been done in RMB, not in dollars.   

The transfer of $60b in reserves to three state banks is small relative to the total recapitalization (see Ma).   Plus, the Chinese government is reported to have promised the banks that it would make up for any currency losses on the dollars that the central bank gave them.   That promise effectively turns the dollars into RMB – just in ways that don’t show up formally.

So how did China recapitalize its banks the last time around, if it didn’t do it primarily by giving the banks its reserves? 

First, lots of bad loans were shifted to the asset management companies (AMCs).  It works like this:  the banks give up their bad loans (at par) and get an offsetting bond from the AMCs.   The AMCs are left with the bad loan.  That bad loan isn’t worth much.  The AMCs end up with a promise to pay say 100 RMB to the banks, and an asset worth around 20 RMB.   That only works if the government effectively promises to make up the difference and pay the AMCs bonds … so in practice, the banks have swapped a bad loan for a government bond.

Second, some bad loans were purchased by the People’s Bank at par.  The banks get cash, which they can use to make new loans.   The PBOC gets a dud loan, which it likely then gives to the AMCs.  The PBoC can afford to take the loss associated with buying a bad loan at par because it makes a lot of money on a flow basis …. 

Finally, the government can just give the banks a government bond.   

It usually works like this.  The banks write down their bad loans.  That creates a hole in their balance sheet – their liabilities are unchanged, but their assets have shrunk.  To make up for the “hole” the government just gives the banks a government bond.  Usually a government bond denominated in the local currency. 

Now, China didn’t always give the banks a domestic government bond – a promise to pay RMB in the future.  It also gave the banks some of China’s reserves – effectively, US government bonds.  

The government gives the banks a US Treasury bond and a promise to swap the payments on that bond for RMB at the current exchange rate, it effectively becomes a Chinese government bond.

That’s the rub.  Chinese banks won’t need dollars if their current loans go bad.  They will need RMB.   And that likely means that the government will end up issuing a lot of RMB bonds.  It just may hide those costs – whether by issuing AMC bonds that everyone assumes the government will have to ultimately pay or by giving the banks dollar-denominated treasuries and an implicit contract (a bond, effectively) to swap dollar payments for RMB at the current exchange rate.

To sum up, dollar reserves can be useful in a domestic banking crisis. 

They are useful if the banking system takes in dollar deposits.   That describes Lebanon, Turkey, Argentina (before 2001) and a host of other countries.   But not China.  China’s banks operate in RMB. 

Dollar reserves can be useful if a loss of confidence in the domestic banking system leans to a surge in demand for foreign currency.

But China has capital controls to limit that risk. 

And even if the controls leak, China’s huge current account surplus almost certainly will generate almost all the hard currency China would need to meet a surge in demand for hard currency from its citizens.   And remember, the conditions that give rise to a banking crisis would likely push China’s current account surplus up to $300b.  

That is a big.   Almost certainly big enough that the Chinese would find themselves adding to their reserves – buying dollars – to keep their currency from appreciating — not running down their reserves to cover capital outflows in excess of China’s current account surplus. 

Of course, if China doesn’t need its reserves to cover a potential capital outflow, it might consider   giving the banks some of the central banks’ extra dollars – really dollar-denominated US bonds —  rather than having the government issue RMB denominated bonds to make the banks whole. 

But the banks still need RMB to be matched, not dollars.  So giving the banks dollars only works if China effectively lets the banks swap out of dollars … 

It turns out that giving the banks a dollar bond plus a swap that protects the banks from exchange rate risk works out to be the same as issuing RMB denominated bonds to recapitalize the banks, holding dollar reserves, and using the interest on those dollar reserves to help pay the interest on the RMB reserves.     

All this is kind of complicated.   It is balance sheet analysis 401, not balance sheet analysis 101.  But I am pretty sure the PBoC gets it.

Think of it this way.   Swapping a dud loan for a depreciating currency generally isn’t a good idea.  Both imply holding assets that won’t hold their value of over time. 


  • Posted by Anonymous

    Thanks Brad for this explanation. Very interesting.

  • Posted by Guest

    (Brad states):”All this is kind of complicated. It is balance sheet analysis 401, not balance sheet analysis 101. But I am pretty sure the PBoC gets it.”

    (Reply): It’s only “complicated” because of the fraudulent monetary scheme practiced by this communist country (which, ahem, happens to be IDENTICAL to the monetary sheme practiced in the SUPPOSED “free” U.S.). That scheme of course is fiat currency, fractional reserve lending and central banking. let’s go back to “balance sheet 101 analysis”. Or better still, let’s call it “government and banking cartel power-grab-crime 101 analysis”.

    The first crime is the issuance of a fiat currency by a government (or through their lackey, a central bank), not backed by any rare specie, such as–you guessed it–gold!!! (Or even silver.) This ability to issue currency (and compel it’s use by force) concentrates too much power in the hands of the government. Again, I stress that COMMUNIST countries can be forgiven for such a power-grab, but there is NO EXCUSE for supposed “free-market economies” to do so.

    The next, even greater crime, is of course fractional-reserve lending. It never ceases to amaze me that even big-brained guys like Brad (and I mean that sincerely, NOT sarcastically), simply go along with this fraudulent scheme whereby people’s hard earned money–money they plae in TRUST at a bank to hold for them–is loaned right back out and NOT available as the DEMAND DEPOSIT THAT THE BANK PROMISED TO THE DEPOSITOR!!! This is fraudulent, inflationary and should be illegal. As an analogy, what if you placed your valuable items in a storage locker and later when you went back to retrieve YOUR property, the owner of the storage company said “oops, sorry, I lent out your goods (with interest attached for me!) to someone else. You’ll have to wait until the other guy gives the goods back before I can give them to you.

    You would be hoppin’ mad and would probably call the police. Yet people have been absolutely BRAINWASHED that, somehow, it is “different” when it comes to their money.

    Finally, the scheme is fully institutionalized with the creation of a “Central bank” to allow the state to further concetrate power in their hands. The central bank serves as “lender of last resort” as pointed out in Brad’s writings above, allowing for the infinite create of ever more credit, debt, mal-investment, booms, busts, pain and suffering for the masses and enrichment of the state and it’s cronies.

    Finally, I find it no less than HILARIOUS that Brad refers to the U.S. dollar as a “Hard Currency”!!! The dollar is nothing more than the largest, most fraudulent FIAT fiction ever dievised by mankind!!! Again, forgive me for belaboring the point, but can anyone reconcile in their mind how a supposedly “free” country and an overtly COMMUNIST country can both practice IDENTICAL monetary schemes?

    As I stated in a previous post, China’s next bust is fore-ordained by their participation in not only the above-outlined fraudulent monetary scheme, but also their participation in the “Dollar Standard” hegemonic scheme whereby China trades REAL goods for UNREAL electronic fiat credits.

    I long for the day that truly brilliant economists such as Brad or Dr. Roubini throw off their yoke of mainstream economics doctrine and embrace a truly free system of money, no fractional reserve lending, and most definitely, no central banking.

    And don’t even get me started on “securitization” and “derivatives”, two additional levels on this great, fraudulent, inverted “pyrimad of fraud and deceit”…

  • Posted by Phil

    I’m not sure what it does to your argument, Brad. But China does have a fairly large amount of foreign currency bank deposits, about $161 billion at the end of the first half of the year, which was down on last year’s total as people have been switching in yuan accounts in anticipation of currency gains. You may remember they raised the interest rate on foreign currency deposits (which also get better tax treatment) last year to encourage more investors to put their money there as the anticipation of yuan gains rose.

  • Posted by Steve Kyle

    It sounds like what you are saying is they will sterilize ex post rather than ex ante – issuing new domestic debt later to fund recapitalization rather than now to fully sterilize if I understood you correctly. Maybe so, but there will be some problems with doing that including but limited to the fact that interest rates later will be lots higher than now and recapitalizing may be a pipe dream for some period of time if balance sheets are entirely frozen with big lumps of worthless real estate on them. Typical outcome in such situations is extended periods of frozen up markets where banks cant mark their book to market and cant make money. Waiting for the market to eventually come back is what eventually bails things out in modern economy real estate busts but it might be a long long wait in some parts of China. Useful conventional wisdom in such booms – it is the later loans in the cycle that go bad the fastest. Which those are I have no clue but I would be ditching any assets connected to such loans starting yesterday

  • Posted by Dave Chiang

    “All is not well between the U.S. and China and, by extension, economies relying on their growth. Neither wants to take responsibility for its own imbalances. Both nations are engaged in a dangerous game of musical chairs. There won’t be enough chairs to go around and the music may stop at any moment.

    If the U.S. wants to lead the global economy, it must do so by example. That means reducing the U.S. trade and budget deficits, and fast. Once that happens, China will have few excuses to delay change in its own backyard. Sadly, it’s not happening, and the world is a riskier place because of it.

    That was Lawrence Summers’s take on things in Singapore. When I asked the former Treasury secretary whether he thought we’d seen the worst of global imbalances, he said: “No, I don’t see evidence that there are strong correcting forces. I think the greatest area of risk is yet to come.”

    Treasury chief from 1999 to 2001, Summers says it’s irresponsible for the U.S. to demand that China tackle its imbalances without addressing its own. Summers also wonders about the precariousness of a rich country like the U.S. being supported by money from developing nations.”

    By William Pesek

    Now that Larry Summers is no longer working in the Clinton Administration or the US Treasury controlled IMF, he can finally speak his mind and be free to be honest for once. – DC

  • Posted by Steve Waldman

    Brad, I think this is a deja vu… I’ll just note that fixing bank balance sheets with RMB-bonds is not a currency mismatch, but it is a maturity mismatch for the banks. Banks that have gotten to a point of crisis will need liquidity as well: they’ll have to be able to sell the bonds for RMB cash, or at least borrow cash against them, to assure depositors that demands for cash can be met. So either taxpayers are gonna have to take a hit and actually fund short-maturity bonds (or zero-maturity cash) issues to banks, or someone has to be willing to buy the long-term IOU issued by the gov’t. Tax raises are a hard sell in the midst of a crisis. Since the domestic, private sector is on net fleeing banks, the only entity likely to lend to banks against gov’t bonds or buy them outright is PBoC. But if treasury issues a bond, and PBoC buys, the net result is PBoC cash fixes broken balance sheets. That’s called monetizing the bad debt. And that is what happens, behind all the AMCs and other crap.

    But when PBoC issues lots of cash to make whole banks that lent to state firms and well connected insiders who are never required to pay, that might be inflationary (if the real economy is in the doldrums, as is usually the case in times of bank panics). People lose faith in the currency. Or they might, but for a trillion dollars in reserve, and a credible peg to the world’s reserve currency. (The US dollar is like Elvis in the Seventies, fat, woozy, drugged, and garish, but still undisputably the King.) That’s why it’s nice to start with $1 trillion USD in the bank and a currency widely acknowledged to be underpriced. It gives a central bank wiggle room to print a lot of money and forgive a lot of sins without provoking a panic, or if the panic comes, $1 trillion is probably enough to ride it out. Heck, if they print enough money, they’ll finally be able to float without having to write down reserve valuations.

    You describe all of this above: that reserves are useful to finance currency holder outflows in a time of reduced confidence, and that China will essentially manufacture local money to resolve a banking crisis, but you don’t so much make the connection that doing the latter may provoke the former. I agree with you that China’s currency/reserve policy might have helped get them into this mess, but if I were a central banker oiling the presses to bail out faltering banks, I’d much rather have 1 trillion dollars in the vaults than not.

  • Posted by HZ

    The banks are not completely state owned any more. Another round of bail-out on non-performing loans won’t be as tenable as the last round, politically. The foreign investers are supposed to help the banks stop making politically motivated loans, because their stakes are on the line as well.

  • Posted by Guest

    Given some of the rants above, thought this was interesting: “…”If we want to be a part of the contemporary world, we need to learn the rules of this world. We can talk as much as we want about Russia’s special way, but it is impossible to determine the special way unless you know what are the rules others are playing by,” he said…”

    Brad – interested in your definition of ‘hard money’. One that came up first on a quick search was: “Funding by a government or organization that is repetitive… A government that uses a hard money policy backs the value of the currency it uses with a hard, tangible and lasting material that will retain its relative value over time.”

    Couldn’t that include trillions of dollars worth of real estate, other commodities… The relative value of all of it dependent on how it is expressed in terms of a fiat currency? Perhaps it comes down to differences in the definition of ‘currency’ and ‘money’.

  • Posted by Joseph Wang

    The trouble I have with a lot of this speculation is that real estate makes up only about 15% of total bank loans, which means that every single real estate loan in China would have to go bad, and you still won’t have as bad a situation as the banks had in the 1990’s.

    The 1990 banking crisis was caused by banks using bad loans to do social welfare spending, and that lending no longer exists. I just don’t see even in some extreme situations, how popping the real estate bubble is going to cause a systematic banking problem. The numbers just don’t work out.

    Also, I’m going to do research on this, but I suspect that the vast amount of lending to SOE’s consists of relatively short term working capital loans rather than long term investment loans, which means that looking at aggregrate credit could be very misleading.

  • Posted by Joseph Wang

    The other issue that I have is that I suspect that the financing for these projects is not coming from bank loans. Local governments can’t easily borrow money, so the financing may be coming from tax revenues or from profits from state owned companies.

  • Posted by Joseph Wang

    I just don’t see the numbers that would get you another banking crisis. Most of the articles that suggest that there will be are “well I see a property bubble here, and so it will destroy the economy.”

    In the 1990’s, Chinese banks issued about $200-400 billion in bad loans to fund social welfare payments. I can’t figure out how you can get anywhere close to that sort of problem with bad real estate and infrastructure loans. I can easily imagine a $50 billion problem developing and a few failed banks, but not a $500 billion problem, and a $50 billion problem is something that doesn’t cause me to lose much sleep at night.

  • Posted by xgbg

    Following Brad’s logic I can’t help reaching the conclusion that “China’s $1 trillion ounce of gold reserve, suppose it has, would be unlikely to be of much use in a banking crisis.”

  • Posted by Ani M

    Even if the bailout needs are higher than the last round of PBoC infusions, they could still come out ok.

    Going by Kyle’s and Walden’s rationale for the need of liquidity and maturity matching, if the GoC issued short-dated RMB paper (say CDs), to account for gap b/w par versus instrinsic value of NPAs transferred to AMCs. Then it could -at least in theory- become an above the line item requiring financing on the GoC’s books (if not a memo item). But since they are short-dated & if they mature quickly and rolled over constantly by the PBoC, while maturities are gradually extended at steadily lower rates then why should it be fiscally problematic? Ofcourse it depends ultimately on the size of bailouts, and the GoC remains very, very sensitive about the true size of NPAs, remember E&Y episode?. As i see it, and subject to the true size of China’s NPLs, it could ultimately turn out to be less of a stock issue and more flow issue that could be repeatedly refi’d by PBoC. From a confidence standpoint, for any such financing to ultimately succeed, the amount of the bailout would have to be less than a year’s worth incremental increase in FX Reserves (to avoid having to disclose to the public any net write-down in FX Reserves due to domestic banking problems). And if, as you say, in the event of a Chinese re-balancing the CA surplus gets much higher (>300Bln) then the consequent increase in FXRsvs and incremental amts. avl for bailout financing would be implicitly higher as well. So where’s the problem?

    Ani M.

  • Posted by Charles

    Oh, that we could have the problems China does. Too much money is flowing into the country, causing the central bank to expand the domestic money supply and keep interest rates low. In a country in which development is desperately needed to raise productivity, there is investment. Too much!

    How about this as a solution to this knotty problem: a corporate income tax?

    There are other means. For example, government employees could be paid partly in dollars (granted, probably at a discount). True, that squeezes them a bit at present, but presumably they get the money later.

    I grant you, your balance sheet expertise puts mine in the shade. But as the saying goes, “Rich or poor, it’s good to have money.”

    Charles of MercuryRising

  • Posted by DOR

    Very lucid analysis, as usual. We’re in danger of becoming used to Brad’s ability to slash through the weeds to revel the path below.

    Still, I have been telling people for a couple of years that China will have a banking crisis; no rapidly rising emerging economy ever gets away without one. And, that it will be a domestic crisis, rather than the usual balance of payments crisis that is so devastating.

    The questions is if, not when. Sadly, the only answer I ever hear from people in China is “after 2008”, as if the Summer Olympic Games have the magical ability to prevent a run on banks.

  • Posted by Joseph Wang

    There are banking crises and BANKING CRISES. I have no doubt that at some point one or several banks will get overextended perhaps due to real estate, and there will be nice pictures of depositors in line waiting for their money, etc. etc. Also, the rural credit cooperatives are a big mess right now.

    But as long as it doesn’t kill the Chinese economy and bring down the government, it’s not something I’m losing sleep over.

    The difference between a $50 billion problem and a $500 billion problem is that in the case of a $50 billion problem, the government can do “tough love” and let the offending bank(s) fail, bail out the small depositors, let the big depositors sink, and let this be a warning to everyone else not to do stupid things. If the problem is large enough to kill the whole economy, then it can’t do that. The threshold is about $100 billion, and I just can’t put together an plausible scenario in which China will face a new $100 billion NPL problem.

    The big problem wasn’t the size of the NPL’s, it was that the NPL’s were due to state-owned enterprises that were massively losing money because they were tasked with social welfare benefits (think General Motors and the airlines). That flow problem was stopped around 2000. Most SOE’s are now profitable. They’ve transferred their social welfare obligations to the government, and the hopeless ones have been shut down.

  • Posted by bsetser

    First, let me note that I am continually impressed by the quality of the comments. I am sure my responses don’t do them justice.

    Second, DC — I would say in general Larry Summers has little trouble speaking his mind. If you look at his history, he has gotten into trouble by saying too much rather than too little. I don’t think he “changed.” I do think he is geuinely worried. I also think (know actually, based on my time at the Treasury) that he deeply believes that the flow of capital should be from rich to poor and from the aging to the young — that basic belief underpinned much of his response to the EM crises of the 90s (and yes, I know you don’t like much of that response). Consequently, I suspect he is more troubled than most by the uphill flow of capital.

    Third, Phil — $160b is a lot. But it is actually a relatively small number for an emerging market of china’s size. Domestic RMB deposits must be well above $3 trillion (I have the data but not on me). $3 trillion is actually conservative. one defining characteristic of China is that its financial system (and household assets) are dominated by the banks — and that most of those assets are in RMB.

    Fourth, JWang — your estimates for the likely NPLs from the current lending boom seem too low to me. Property is only @ 15% — tis true, but i think it is higher for a couple of the big SCBs. (And i forget whether property exposure includes both mortgages — i.e. household credit — and lending to developers). And the banks aren’t financing social safety nets. But you can loose moeny financing over-investment in real assets too — as well as white elephants favored by the provincial governor. I am not as worried about a repeat of the 90s style bad loans from the Iron Rice bowl firms as a Chinese version of the Asian crisis, stemming from over-investment (fueled by connected not just directed lending). My paper has my ballpark estimates — which assumed some fraction of new loans extended since 2002 will go bad. I am with DOR on this: most emerging market banking systems that ramp up lending really fast tend to make mistakes.

    Fifth, yes, issuing long-term bonds to the banks creates a maturity mismatch (so too does an NPL … it is illiquid and de facto long maturity in addition to being impaired ). Sometimes recap bonds carry a floating rate to help the banks with the resulting interest rate risk. But I don’t think this implies that the banks need to be recapped with short-term government bonds.

    This is how it usually works —

    The government gives the banks a long-term bond (or guarantees a bond issued by an government owned AMC). that provides the banks with a performing asset — and something that can serve as collateral.

    IF the banks are short on cash, the banks can post the bond as collateral for a loan (i.e. cash) from the central bank …

    basically the government provides the good asset that central bank needs as collateral, and the central bank provides the liquidity the bank needs to meet withdrawals.

    Yes, that means increasing the money supply — but in a banking crisis, demand for cash can rise.

    When to dollar/ euro reserves come in useful — if residents don’t want to hold RMB cash and instead want dollars …

    effectively that adds another step. the central bank supplies the banks with cash (in local currency), and the depositors trade the cash back to the central bank for the central bank’s foreign exchange reserves.

    The core issue is whether demand for dollars woudl surge from less than zero to over $250b (enough to offset the huge chinese current account surplus) should China’s economy slow/ many recent loans start to go bad.

    In general, I don’t think so.

    Finally, yes, there is something of an irony in the conventional characterization of the $ as a hard currency given the United States external imbalance. I think i have made a similar point in the past, even i lapsed here …

  • Posted by Joseph Wang

    I wouldn’t be too surprised if a collapse in the real estate market causes one or more JSCB’s to collapse, but that can be handled by asset securitization to spread risk and by depositor insurance. Real estate bubbles have happened before and banks have collapsed before in China.

    However, I don’t think that this is the big problem in the banking system. The big problem is the fact that the Agricultural Bank of China is a mess, the rural credit cooperatives are basically broke, and this is all related to the fact that the rural interior isn’t growing very much. The big problem with the “white elephants” that are being built is that they are eating up capital that could be used to increase standards of living in the interior. Credit is easy to get for projects in the coastal provinces, but practically impossible to get in the interior.

    I think this is part of the general issue is that if you are rich, you find it easier to get credit which means that credit is flowing from the poor provinces to the rich provinces to the United States.

    One consequence of this is that I just can’t see overinvestment being a serious problem in the next decade since that can be dealt with by an economic stimulus that boosts consumption in the interior provinces.

    I’m also not worried about a repeat of the Asian crisis in China, people have spent a lot of time and effort trying to figure out how to prevent this, and the general rule is that things that people spend a lot of time worrying about get resolved. It’s the things that people don’t worry about that bites you.

    What I am worried about is a crisis that involves something that no one has thought of. I’m even more worried about a “non-crisis” in which things get progressively worse and worse, but in which there is no moment in which people notice that things have gone seriously wrong.

    Right now, the scenario I’m thinking about involves the question, what are the financial implications if around 2012 Iraq totally collapses and the United States is forced to withdraw, or what happens if there is a confrontation over Iran?

    Note to D. Chiang: Something that you have to realize is that regardless of why, if US global power collapses it will pull China down with it.

    One of the very curious Chinese papers that I’ve read argues that it is not in China’s interest for US power to decline too rapidly. I can easily imagine a situation in which China is very actively trying to prop up US power (and one can argue that this is already happening).

  • Posted by HZ

    Is it possible that banking crisis is not so big an issue as you make it out to be in an economy where the banks (state owned ones anyway) don’t function all that well in the first place?
    Here is my reasoning:
    1) From accounting point of view financial asset and liability denominated in RMB always match on a national level. Bad debt only exists if you take the view from the banking sector only. A bad debt for the state banks means someone else owns the capital now – an argubaly good thing in China since it transfers capital to the private sector. Since state banks are not that great at allocating capital anyway the odds are that capital efficiency goes up with the bad debt. This is not to deny the inefficiency that led to bad debts in the first place but the reinvestments of these runaway captital are potentially a lot more efficient.
    2) Depositors are implicitly protected by the state so there is no liquidity squeeze. There is no need or possibility of one anyway for a fiat money.
    3) If the state banks go under the best way for savage is to allow private banks to start up and take over the business of capital allocation. As stated in point 1 there will be no shortage of capital — just that the state banks will have squandered them. Although I am doubtful that the Chinese leaders will see it this way, this is a good way to have a truely functional financial system.

    There had been for years a lot of predictions of doom to come because of the frail banking system. They never bore out and I think there is a reason why that is the case. It is not optimal but it is self-limiting: (state) banks are not good capital allocators => economy is not dependent on their capital allocating function to suceed => their failure may be a good thing and the sooner the less damage done.

    I think this chain of logic works for the 90s. Whether it still holds I am not sure. The state banks are hugely capitalized and waste of so much capital could do more harm than I supposed.

  • Posted by bsetser

    HZ — well, some of the predictions of doom from the frail banking system did come true. the cost to taxpayers of moving the bad 90s loans off the balance sheet will be very significant. China simply deferred taking those costs — and in the meantime its economy grew dramatically, so as a % of GDP, they will seem smaller than they would have been as a % of gDP had the full costs been incurred in say 02.

    the fact that deposits are protected (or assumed to be protected) is a key reason why large NPL losses haven’t led to anything more serious than losses to the taxpayers — particularly in a context where the bank deposit base is growing fast so there is no shortage of capacity for new financing.

    as for the argument that bad loans are just one way of transferring capital from the bad state sector to the good private sector, I am not convinced. The “good” private sector is presumably the set of borrowers who take a loan and repay. the private sector that takes a loan and doesn’t repay could quite easily have squandered the funds. And I rather suspect that a lot of borrowers in that position tend to get financing because of close ties to the local party (probably especially true for property developers … ).

    Does the lending a stimulus to activity, even if not all the investment it finances has a high return — absolutely.

    It is a covert way of transferring savings to the private sector? I am not so sure. And transfers determined by an unwillingness to pay hardly seems likely to transfer savings to the folks most able to use it efficiently.

  • Posted by Ani M.

    >>A bad debt for the state banks means someone else owns the capital now – an argubaly good thing in China since it transfers capital to the private sector. Since state banks are not that great at allocating capital anyway the odds are that capital efficiency goes up with the bad debt.Capital?) and the reverse holds true for the rest of the private sector? Or are you getting at something else?

    Ani M.

  • Posted by Ani M

    Sorry, ignore my last msg. for some odd reason all my comments didn’t post. Thx,

  • Posted by Joseph Wang

    Crisis -> uncertainty -> inability to plan -> generally bad thing

    The thing about the state banks is that they are places where people park their cash, so that you can do financial transactions without carrying suitcases of cash. They aren’t terribly good allocators of capital, but they provide a safe place for people to hold their money so that they can use it to start their own businesses or finance the business of friends and families.

    The bad debts were in the hands of state-owned enterprises, so allowing a crisis and wiping those debts, actually moves things in the wrong direction. It would have wiped out depositors who are using the cash for economic growth, and it wouldn’t have done anything to create a healthy economy.

    Also, I don’t think that state/private ownership makes a big difference. What matters is that the enterprise is profit-drive rather than revenue-driven.

    Finally, do not underestimate the difficulty of creating institutions. The problem (and you see this in Iraq) is that if the old institutions fall, there is generally chaos, and when there is chaos, then people aren’t in the mood to create new institutions. A major bank is just not something you can create overnight.

    The reason the predictions of doom and gloom didn’t happen is that people looked at the predictions, spent a *huge* amount of effort try to fixing the problems, and fixed them.

  • Posted by Joseph Wang

    One thing that is also missed is how far China has come, and how difficult the process of economic transition was.

    In 1978, there were no real banks, and all financial institutions and factories were government ministries. Banks in China today may be rather inefficient capital allocators, but in 1978, there wasn’t even the concept of capital allocation.

    In a command economy, there is no concept of banking, capital, debt, prices, or even money since factories just act according to central orders. This poses a big problem in that without prices, there is no real way of “keeping score” to see whether what you are doing makes sense or not.

    Getting from that point to where we are today was phenonmenon, but the difficulty in actually creating new institutions to handle all of this is why I don’t subscribe to arguments of “lets destroy the old system and the new system will magically appear.” It takes years for the new system to appear.

    In China, what happened was that the old system and the new system existed side by side for about two decades, which gave new system time to mature while not causing things to collapse. In Russia, the old system collapsed, there was no new system, and you ended up with total chaos.

    In the case of state factories, the fact that you had a private sector generating jobs, meant that you could slowly shut down the factories that weren’t profitable, and those that were could participate in the new system.

    This actually gives you a timetable for Iraq. To actually do anything useful in the Iraqi economy is going to take about 20 years or so.

  • Posted by xgbg

    I would bet one dollar that any financial crisis China may suffer in the future will be coincided with if not proceeded by a crisis in the United States. Comparing bank A holding a NPL with bank B holding treasury securities, you may take it for granted that B is in a better shape. But what if the US government issued the debt for consumption? Is it some kind of “performing loan”? If the crisis is about financial assets, the Chinese government can always print money. If it is about goods and services, the trillion dollar reserve is a nightmare for the US as well as the entire world.

  • Posted by Emmanuel

    I am terribly late to this post. Anyway, the WSJ article struck me as a principal-agent problem between central government and the provinces:

    Compounding the problem is the way local governments are forced to fund their investments. China’s central government doesn’t allow them to raise local taxes or issue debt. Instead, they are cashing in their most valuable asset — land. The sale of land now accounts for 40% to 60% of all local government revenue, according to Ms. Wang of the Chinese Academy of Social Sciences.

    Will devolving responsibility for raising taxes and issuing munis to the provinces fix things or make them worse? On one hand, giving provinces more leeway in fiscal matters should make them more responsible with their projects. OTOH, the provinces have spotty records when it comes to fiscal management, so this cure might be even worse than the problem. And then there’s the question of who’d buy their munis in the first place…

  • Posted by bsetser

    Emmanuel — good points.

  • Posted by HZ

    “It is a covert way of transferring savings to the private sector? I am not so sure. And transfers determined by an unwillingness to pay hardly seems likely to transfer savings to the folks most able to use it efficiently.”

    I certainly agree that the problem loans should be assumed to have been inefficiently used – the hypothesis is that the follow on reinvestment of the capital will be more efficient. Herein lies the contradiction: an official/businessman cavalier with squandering public money may well be wise in managing their own private funds.

    Let’s trace how the capital circulates with NPL. Let’s say an official wants an image building done — and let’s say there is now way to rent out the building to recoup the investment, with negative ROIC, so the state entity that takes out the loan and owns the building will eventually go kaput. But the money does get spent on building materials and construction, transferring capital to the materials/building industry. Yes it is preferrable if the building had turned out to be a commercial success were the initial investment to be wise, nevertheless this stimulated construction activity and allowed (potentially) private players in the related industry to build up capital. The initial process may have been terribly inefficient but the private parties may be assumed to seek better ROIC now that they own the capital. Granted this process should not be allowed to happen over and over again but at the time when the private sector was starved of capital it might have proved positive overall or at least not as negative as it seemed. (Other factors may mitigate the situation too, initially empty roads may quickly get filled as more people own cars and empty buildings may eventually get used as urbanization progresses etc.) An outright fair privatization scheme would have been a lot better but then again we would all like to live in Eden.

    I think an analogy can be made to the boom-bust cycles in industrial economies of the bygone era — sure the capital efficiency were hurt by the booms but it sure beated the slow growth of the agrarian peers.

    Hopefully the above clarified my arguments. I agree with many things you said.

  • Posted by Joseph Wang

    Emmanuel: Getting together a rational taxing and borrowing system together was a big struggle in the 1990’s. The way the system works now (after trying about a dozen things that didn’t work) is that the bonds for local development are issued by the China Development Bank and the money then sent to local municipalities. One thing that really, really doesn’t work is to give local municipalities authority to borrow.

    HZ: A lot depends on the contents of the NPL’s. The NPL’s of the 1990’s were spent basically on unemployment benefits for laid off workers. A factory is broke. It needs to lay off workers, but if those workers go out on the street with no job and no welfare, they will riot. So this caused money to be dumped from the banks to the SOE’s.

    The trouble with your scenario happens when the person who lent the money for that building wants their money back. People that go to the bank expecting to get their money, and not getting it, get angry.

  • Posted by HZ

    You are describing a liquidity squeeze. You know that couldn’t really happen.

  • Posted by HZ

    Let me recast the argument this way:
    – With a profit maximizing private sector and a profit insensitive state sector eventually profit accrues to the private sector (and cycle back through taxation).
    – NPL was a way to inject liquidity. It was a bad bad way to do so but it was still a lot better than no liquidity. Whether it was welfare or not it eventually accrued to the profit seeking private enterprises.
    – NPL is not good. Banks should not be in the business of sustaining failing state companies. But functioning banks were not the reason that the economy performed well (the economy performed well despite the weak banks) therefore the downfall of state banks would not necessarily lead the economy down the drain — assuming liquidity is wisely maintained.

  • Posted by Joseph Wang


    The problem with looking at the Chinese economy is that economic models are based on assumptions and in a half transitioning economy, those assumptions may not be valid. There was a *huge* amount of trial and error to figure out what works and what doesn’t, and a system based on trial and error implies no crises. You just can’t experiment when things are falling apart around you.

    Part of the problem in these discussions is that I don’t *know* what would have happened if something different went on. I just base things on what I see, and what I’ve seen is that crises are bad, and provoking a crisis based on a economic and political theory on what should happen is unwise.

    The big problem is that your economic and political theory is likely to be wrong, which means that you are going to be in a position of trying to come up with Plan B at the worst possible time. (Witness Iraq and Russia). All economic and political plans need to have the question “well suppose you are wrong about X, what happens?” asked.

    Problem with the first section. Part of the challenge of the 1990’s was to set up a taxation system. You couldn’t set up a social welfare system or a good private economy without a system of taxation and that took about five years to put in place. The reason that NPL’s where used for social welfare, was that there was no other system in place in 1992.

    Problem with the third section. For the private economy to work, there had to be two conditions. First, the private economy had not to be overwhelmed by demands for employment before it was ready. Second, the private economy needs a safe place to park money. The state banks in the 1990’s filled both those roles.

    The other thing that happened in alot in 1990’s, was that you had one person in the family working in an SOE and getting the social benefits of the SOE (a cheap apartment, daycare, cheap health care, a steady income) which allowed someone else in the family to risk going into the private sector. That cushioned the blow in the late-1990’s, since when the SOE’s were shut down, they had other business to support themselves. If the SOE’s where shut down in the early 1990’s, this wouldn’t have happened.

    The basic problem with “let it fall part” is that it takes a while to build an institution. Three to five years. That might not sound like much (and it really isn’t), but it’s a long time to be standing in front of a teller waiting for your money.

    The other assumption (which I think is incorrect) is that private institutions are inherently better allocators of capital than public ones. What seems pretty obvious now (and it wasn’t obvious in 1990) is that who owns the company is less important than what the goals of the company are. Three of big-four state owned banks are now in pretty good shape because even though the Chinese government is the main owner, the banks have been ordered to “make a profit” and the rules are such so that to make a profit they have to do things like increase efficiency, which they’ve done.

    The problem right now is that the SOE’s are making so much profit that no one knows what to do with it. It’s a better problem than NPL’s, but just like figuring what to do with the epidemic of obesity and diabetes that comes with lots of food, it’s still a problem.

    What happened in Russia was that a botched privatization caused enterprises to end up in the hands of the politically connected and the motive there was to loot the company.

  • Posted by Joseph Wang

    HZ: Thinking about it some more, what you are saying may make some sense when are you talking about the smaller locally owned banks. In the case of the big four, a bank failure would create huge amount of economic problems, but the argument that NPL’s aren’t so bad may make more sense at for the smaller banks that wouldn’t kill the economy if they failed.

    One of the great decisions that the PRC government has made is to let smaller banks fail when they do something really stupid. (GITIC for example).

  • Posted by HZ

    Thanks for your thoughts. I think there are two rather separate aspects to the banking system: one is transactional, and let’s include working cap credits in this category; one is capital allocation where NPLs eat into the bank equity capital. My argument would be that there is no reason why failure of the second part must lead to the failure of the first part. But otherwise I agree with you — no grand plan for change. I am just trying to get a sense of how a banking crisis might unfold and to me it in all likelihood it won’t be a huge deal. Just the nature of the beast: inefficiency actually means that the banks are not highly leveraged and the economy is not totally dependent of the capital allocation prowess of the banks.

  • Posted by Guest

    Why would The Bank Of Of America buy NPLs from other countries. To me these loans are just useless paper.