Brad Setser

Follow the Money

Cross border flows, with a bit of macroeconomics

Print Print Cite Cite
Style: MLA APA Chicago Close

loading...

Just how much money does China have? How fast are China’s foreign assets growing? And how much is hot money?

by Brad Setser
July 21, 2008

Answering these questions has been a long-term obsession of mine. I am not sure I have the answers, even now. China’s government doesn’t make tracking the growth of its foreign assets easy.

But I do have fairly detailed estimates. What’s more, these estimates are generally based on data that China itself releases, often in somewhat obscure places – though, given data lags, I sometimes have made estimates to bring a data series up to date. Of course, both my interpretation of the data and the assumptions I have used to produce data through the end of June could be off. These are estimates.

Based on the assumptions laid out in the technical notes, including an assumption that the transfer of foreign exchange to the CIC was largely completed by the end of q1 2008, I estimate that China’s government currently manages between $2.3 and $2.4 trillion in foreign assets. The central bank (SAFE) manages $1.8 trillion, the CIC manages $109 billion (my assumption), and the state banks manage $430 billion. This implies that China’s state banks have become one of the largest reserve managers in the world — their combined portfolio trails only the reserves of Japan and Russia. The combined portfolio of the CIC and the state banks it formally owns would make it the second largest SWF in the world.

chinese-foreign-assets-h1-08-1.JPG

The pace of growth in China’s foreign assets is equally impressive – I estimate China added, after adjusting for valuation changes, something like $785 billion to its foreign portfolio over the last 12 months (i.e. from June 2007 to June 2008). Just a bit under $420 billion of this comes from the increase in the central bank’s reserves. The CIC got an estimated $107 billion (after adjusting for its purchase of Huijin and the funds it injected into the CDB), and the banks holding of foreign exchange increased by a little more than $250 billion. Most of that increase ($200 billion) comes from the fx held as part of the banks reserve requirement; the remainder comes from an apparent increase in the banks swaps position with the central bank in the second half of 2007.*

chinese-foreign-assets-h1-08-2.JPG

For the first half of 2008, China added – as best as I can tell – about $430 billion to its foreign assets, with $250 billion coming from the increase in China’s reserves (after adjusting for valuation gains), $90 billion coming from the increase in the CIC’s foreign assets and $90 billion coming from the increase in the banks reserve requirement. This estimate is close to the estimate of Michael Pettis and also to the estimate of Logan Wright.

It is worth noting that if this analysis is right, the increase in China’s foreign assets over the last 12 months was larger than the US trade deficit over that time – and far larger than the United States’ non-oil trade deficit.

chinese-foreign-assets-h1-08-3.JPG

Not all of China’s foreign assets are in dollars. But if a significant fraction are, China isn’t just financing its own exports to the US. It is in some sense financing everyone’s exports to the US. Or alternatively, it is financing capital outflows from the US. This has been a quiet bailout only because China hasn’t made its financing of the US conditional on US policy changes. The scale of the financing is enormous.

The data set tracking China’s foreign asset growth also allows me to estimate hot money inflows into China, as the growth in China’s foreign assets can be compared with China’s trade surplus, its estimated interest income, and FDI inflows. Before doing this, though, I want to acknowledge that Qing Wang’s critique of this methodology for estimating hot money flows is right. Summing estimated interest income and trade doesn’t produce a data series that matches the current account balance perfectly, as it leaves out transfers (largely remittances from the diaspora). The monthly FDI data also doesn’t capture all FDI inflows – in recent years, the FDI inflows in the final balance of payments data have been larger than the sum of the monthly FDI inflows released by China’s commerce ministry. And finally, there are inflows – QFII inflows for example – that wouldn’t normally be thought of as hot money.

As a result, this methodology will tend to overstate “true” hot money inflows if the trade and FDI data doesn’t include any disguised inflows. My hope is that the errors offset – that hot money disguised as FDI is large enough to balance the “not hot” flows counted as “hot money” by this methodology. The main virtue of the trade+ interest income + FDI methodology is that it allows a consistent comparison of the gap between known sources of foreign asset growth and observed foreign asset growth over time, not that it provides an accurate measure of hot money flows. It captures trends better than levels.

So what does this data show?

Well, there has been a large increase in the gap between my estimate of China’s foreign asset growth and easily identified sources of reserve growth over the last 12 months – a gap that likely corresponds with a rise in hot money inflows.

chinese-foreign-assets-h1-08-7.JPG

This makes intuitive sense. Over the last 12 months, the US has cut rates while China has held rates constant and the pace of RMB appreciation (against the dollar) has picked up, increasing incentives to hold RMB rather than dollars.

The same data can be examined on a rolling 3m basis to get a better idea of recent trends.

chinese-foreign-assets-h1-08-5.JPG

The recent data suggests that hot money flows have fallen off a bit – as total foreign asset growth in May and especially June was well below the level in April.

The high frequency data also provides ground for caution though. Look at q4 2006 v q1 2007. The data suggests hot money outflows in q4, and then huge inflows in q1 2007 (q1 was the quarter when Chinese reserve growth really popped up). While I have little doubt that there was a big increase in Chinese foreign asset growth in early 2007, I suspect that the transition was a bit smoother than my data series implies – i.e. I am missing some foreign asset growth in q4 2006, and my adjustments may overstate foreign asset growth in q1 2007. That is just a hunch though. In a similar way, I wouldn’t be surprised if the June increase in China’s foreign assets is somewhat larger than the data China just released implies. But that is very much a hunch.

A plot of the gap between foreign asset growth and trade, interest income and FDI inflows – with the gap used a proxy for hot money – shows that inflows have leveled off, or perhaps even turned down recently.

chinese-foreign-assets-h1-08-6.JPG

But they remain at a very elevated level. It still isn’t clear, at least to me, if China’s controls are effective enough to allow China to sustain a significantly tighter monetary policy than in the US while it still manages its currency against the dollar. We will see.

And in the interim, do read Dr. Yu’s overview of China’s economy – and the current monetary policy debate in China.

Technical notes

I have estimated the state banks’ foreign assets using their fx liabilities, not their reported fx assets. I have in effect assumed that they match their liabilities with assets, and fx liabilities with the central bank are held offshore.

The key line items I have used in my estimates include

In the PBoC data on the “Sources and uses of funds in financial institutions (in foreign currency), “other items” (in the set of “sources of foreign exchange”). This line item jumped in early 2004, just after the first round of Huijin recapitalizations. This data series has not been updated for 2008. I have assumed that it hasn’t changed.

And in the same table, the line item under sources labeled “purchases and sales of foreign exchange.” This line item increased rapidly in 2006, when the banks started buying large amounts of foreign debt. This line item fell in the first part of 2007, when reserve growth suddenly accelerated, and then moved up in late 2007 (when reserve growth slowed, relative to the trade surplus). This data series has not been updated for 2008. I have assumed that there were no additional changes.

I also have assumed that the “other foreign assets” reported on the PBoC’s balance sheet reflects the banks fx reserve requirement. Stephen Green, Wang Tao and Logan Wright have all argued that there is a close correspondence between the rise in this line item and the rise in bank reserve requirements likely held in foreign exchange. This data is current through April. Based on Logan Wright’s work, I have assumed a $18b increase in May, and another $36b increase in June.

The result of these adjustments is a data series that is consistent with China’s balance of payments data — if one assumes that Chinese banks have been accumulating foreign assets on behalf of the central bank. We know, for example, that private Chinese investors (likely the state banks) bought a lot of foreign debt in 2006 (see Qing Wang). There also was an increase the state banks fx liabilities from the “purchases and sales of foreign exchange” in 2006, a line item that likely reflects one leg of the banks swap transactions with the central bank. In 2007, private Chinese investors (again, likely the state banks) increased their deposits in the international banking system. There also was an increase in the PBoC’s “other foreign assets” in 2007, a line item that likely reflects the foreign exchange the banks have been asked to hold as part of their reserve requirement.

I have assumed that the CIC’s foreign assets increased by $13 billion in q3 2007 (largely because of purchase of Huijing), by $4b in q4 2007 (largely because of the recapitalization of the CDB and one other state bank) and $90 billion in q1 2008. This estimate is subject to substantial uncertainty, especially with respect to the timing. It is a lower estimate for the CIC’s foreign assets than most because I am trying to capture the foreign assets actually managed by the CIC, not funds shifted to the state banks.

Finally, for the first three line items, I have looked at changes relative to a baseline – i.e. changes relative to the levels of early 2003 for the “purchases and sales” and ‘other liabilities” and changes relative to end 2006 for “other foreign assets.” I wanted to capture the policy changes that I think the “pops” in each of these line items represent, not whatever was going on before. This is a bit arbitrary.

I welcome critiques of my methodology. There is a risk that I am misinterpreting the bank data. And there is also a risk that I am double counting something, missing something or generally off a bit. The methodology is essentially the same as the methodology I used in my January paper on China’s foreign assets, but it has been updated to reflect the banks new fx reserve requirement in a more systematic way.

65 Comments

  • Posted by Twofish

    DC: The Federal Reserve bailout of Wall Street with a “cheap money” monetary policy is rapidly destroying the US Dollar.

    And Bernanke believes (and I agree) that this is less bad than a string of bank failures caused by the tight money policies followed after the crash of 1929. If you go back to 1931, everything you are saying now was said about the economy then, and it turns out in hindsight to have been completely the wrong set of policies.

    So this time, you do something different.

  • Posted by Twofish

    bsetser: A slow, consensus driven policy making process that has been behind the curve — constantly making changes to the XR that are late and too small to matter.

    I don’t think that things are too late or too small. If you have social revolution and mass protests, then yes things are too late or too small. China has moved fast enough to avoid a major crisis, and moving too quickly runs the risk of major crisis occurring.

    As it is Chinese reserves are over $2 trillion. If the peg was still in place, then would be larger.

  • Posted by bsetser

    2fish — I would argue that the scale of the current boom, and the internal tensions it has generated (negative real rates = big risks of over-investment) risk creating a social revolution if the boom turns to a (macro) bust. moreover, if china tried to export its way out, i suspect the international economic system would turn against china — which wanted to rely on global demand on the way up, and on the way down … just a hunch.

  • Posted by Twofish

    bsetser: I would argue that the scale of the current boom, and the internal tensions it has generated (negative real rates = big risks of over-investment) risk creating a social revolution if the boom turns to a (macro) bust.

    I’d argue that China is at such a low state of economic development that overinvestment is not a major concern. Misinvestment is a problem, but overinvestment isn’t.

    One thing to remember is that China has gone through several boom-bust cycles before and only one of them (1989) lead to mass social unrest. The tools that the Chinese government has now are much more advanced than anything that it had in 1993 or 1998 when the last set of busts happened, and they were able to damp down the last cycle in 2003-2004 without causing major trouble.

    bsetser: if china tried to export its way out, i suspect the international economic system would turn against china

    On the other hand boosting domestic demand isn’t a particularly hard thing to do. Fire up the helicopters and drop dollars from the skies.

    bsetser: moreover, if china tried to export its way out, i suspect the international economic system would turn against china — which wanted to rely on global demand on the way up, and on the way down

    I think that this is much less likely than first appears since you have large parts of the international economic system now highly dependent on Chinese exports.

    If you have an economic slowdown that hits employment, at that point you may see the economic system turn against China as it proceeds to tear itself to pieces like in did in the 1930’s, but Bernanke is doing what he can to fire up the helicopters in the United States to make sure that doesn’t happen.

  • Posted by Rien Huizer

    Brad,
    Technical question v had for long time: USD assets on the books of CIC’s banks: are these hedged, and if so (a) where and (b) at what implied rates of interest (USD and RMB). Furthermore, where are they placed? I see a range of posibilities here: (1) not hedging (would be very capital intensive under Basle II, which I believe at least BOC has to adhere to (2) FX forwards need some CNY and some USD rate. (3) a portion of the USD deposits can probably be used internally (lend to customers, fund outstanding in trade finance, fund overseas offices and subs, etc) . (4)If they hedge, they would probably only hedge the surplus over their internal needs (and do that with the POBC?

  • Posted by Twofish

    Huizer: Technical question v had for long time: USD assets on the books of CIC’s banks: are these hedged?

    My understanding is that much of the US dollars held by the banks are matched against US deposits held by the banks so a change in the value of the dollars is matched with the change in bank liabilities. To the extent that this isn’t the case. the PBC has been known to sell banks currency options which means that the currency risk is in the hands of the PBC rather than the banks.

    One other point….

    bsetser: moreover, if china tried to export its way out, i suspect the international economic system would turn against china

    I don’t see the global economic system breaking in this particular way. The reason why is that if the Chinese government felt that boosting exports was the best way of stimulating demand, it would write the US a nice big loan for $500 billion which would be used to buy Chinese goods. I really don’t see the US tearing up this check rather than cashing it.

    Long before the US starts going protectionist, someone in Beijing is going to wonder if it really in China’s interest to keep loaning the US so much money, and that is the limit in the current situation.

    This is also why I think there are limits to how bad the situation in China can get. Suppose I take $50,000 of my own money and invest it in something like Shanghai stocks and the stock market drops 50%. Whoops. I kick myself, but nothing really bad happens.

    Now if I borrow $50,000 and the market drops 5%, I’m really in trouble. If it drops enough then the person that loaned me money could be in trouble. If we are all loaning each other money (i.e. the Japan situation) then we all are in big trouble.

    In the case of the Chinese economy, there is this huge pile of cash. So there will be bumps in the Chinese economy, but since China is a net creditor, nothing seriously bad will happen if the value of its assets drop, as long as someone in China has the cash reserves to absorb the loss. This makes it very different from Latin America and other emerging markets.

    Getting back to the helicopter analogy. If you have a demand problem, then all you have to do is to fly helicopters and drop money from the skies. So why do nations get into such serious trouble…..

    Well, if someone else owns the helicopters……..

  • Posted by bsetser

    the dollar deposits of chinese banks offset the banks domestic dollar loans; actually, that is no longer the case, as dollar loans have been rising faster than dollar deposits. the exact fx position of the state banks is a state secret. The fx the banks have from swaps with the PBoC is dollar-hedged (but i don’t know the precise terms of the hedge). The fx banks the banks received from their recapitalization is rumored to be hedged as well — tho some of the original hedges have by now likely expired, and it isn’t clear if the hedges were rolled over or not. Finally, there are rumors that the banks have been promised (or received) hedges for the fx they now hold as part of their reserve requirement, but that CERTAINLY hasn’t been disclosed.

    Sum it all up and I cannot answer Rien’s question with confidence. These tho are the kinds of questions bank analysts in China should be asking.

    finally, 2fish, bad investment is often correlated with high levels of investment. and i suspect the evidence suggests that even poor countries can get a bit drunk and invest a bit more than makes sense.

  • Posted by Rien Huizer

    Brad,

    thanks, not only the FX positions of Chinese state banks re secret!
    The way I understand it then is that the banks have three sources of USD-related liabilities:

    (1) deposits from customers who have permission to hold FX deposits onshore

    (2) deposits received as part of the recapitaliation (I would assume that these deposits have legal aspect (subordination, posibly convertbility) that allow POBC to treat these deposits as Tier 1 or Tier 2 instruments under the Basle accord

    (3) “deposits” resulting from FX swaps with SAFE or POBC

    All of these could be hedged, either by entering in spot/forward swaps or by using natural hedges, like USD loans to onshore customers, or by placing FX deposits with offshore affiliates, or with POBC.

    Re (1) then, that would look like the most likely source for funding onshore commercial assets (onshore loans, export bills etc). I would guess that the surplus, if any could first be used for placement with own offshore affiliates (who my be subject to stricter arbitrage controls than ordinary western banks) . The remainder would probably go to SAFE. If that is the case it would be logical if the deposits with SAFE would offer the lowest profitmargin (yield-/- cost of funds)

    Re (2) These deposits may carry a different rate(payable) (in FX ) than that paid by POBC/SAFE to the same banks on their surplus FX cash, because they are more risky (assuming again that they have low seniority in order to count as capital) Also, in order to hedge them (otherwise gigantic FX exposure would exist), they probably swap thes depositis with POBC/SAFE by selling he FX spot and buying it forward. That would convert the FX into CNY. In oder to price the forward one needs an interest differential. The question then would be would (from the commercial bank’s perspective) the implied USD rate receivedbe the same as paid on th recap deposit or a different one. Also, would the RMB rate paid be he same als, for instance received on CNY reserve requirements, etc

    Re (3) similar questions as with (2) I am not familiar with the balance sheet structure in FX (doubt anyone is) but if e go back to (1) swaps could of course be used for both investing surplus FX (instead op depositing and hedging into RMB separately, or to fund deficits in the commercial business.

    The point of this is that all these transaction types use interest rates, and that these interest rates may differ from rates used in customer business, and certainly from international (LIBOR) based rates. Hence it is posible that the state subsidizes the banks or vic versa. I would expect that it is the latter and that would make this detailed (apologies) quesion a very legitimate one for equity analysts

  • Posted by Rien Huizer

    I am puting this in separate comment but it accompanies the previous one

    Although the whole system is quite controlled, it is not possible in a vast country like China and commercial banking something that has only been around for a decade at best (remember that in a centrally planned economy banking is a mere bookkeeping and cash dsipensing function) to effectively control (a) a commercial exchange licence system (the banks would have to police that) and (b) enforce the policy governing application approval throughout the country. In addition, the major banks are still completing their first generation back office systems and given the fact that only 4 years ago FX risk was very unusual given the peg, most probably the banks have very little infrastructure, not only to process large numbers of FX transactions, but also to manage the associated credit risks (typically a firm wishing to sell future USD receipts forward would need a credit facility of some kind. I would not expect that the commercil SOE banks would have a good credit mangement system for FX forwards. Posibly during the past year, more and more bright sparks found ways to “hedge” and perhaps, during the past 3-4 months, a lot of that may have matured. If the dministrative situation in the banking system is as I suspect, that might actully not be easy to track centrally and could easily lead to pretty big surprises. The obvious policy response could of course be that all fwd business would require 100% cash collateral in CNY. I wonder if any of these aspects (1) deficient administration especially for credit risk management (2)weaknsses in design, enforcement and approval policy of FX control, have had attention in the Chinese language literature recently.

    One way to make that type of business more selective (shaking out speculation) is to require a high cash deposit
    (a bit similr to a margin on an exchange traded product).
    I wonder if there is something of that kind

  • Posted by Rien Huizer

    sorry, pse ignore the last paragraph of course..

  • Posted by Gregor Neumann

    TwoFish #56: A closer look at the US recession 1918-1921 should be a warning. A rich industrial nation with a trade surplus but severe overcapacity is going to hurt its home market, if overseas markets are overburdened. How long will China be able to absorb excessive inventory, if customers in the US and Europe say “no thanks”.

    You said: “The reason why is that if the Chinese government felt that boosting exports was the best way of stimulating demand, it would write the US a nice big loan for $500 billion which would be used to buy Chinese goods. I really don’t see the US tearing up this check rather than cashing it.“

    This has been going on for too long now. The US is credit market is contracting. Overextended institutions need money to cover losses, but they are currently not generating new business. This will hurt the demand side in the US and therefore the supply side in Europe and China.

    Has China enough cash to buffer job losses? What is the financial situation of the “new middle class” in the hot spots like Shanghai? Will they “just lose savings” or did they use credit to buy real estate and stocks as well? We hear reports of Chinese companies struggling for money. It is hard to imagine that this will not affect their works, who are customers of retailers.

    I worry that China is trying to “grow out of the problem”, because this strategy has worked pretty well. The landscape in its main markets has change. And I fear that this will stall internal demand.

  • Posted by Twofish

    bsetser: the exact fx position of the state banks is a state secret.

    Most of the major banks are all traded on HK exchanges and are subject to the same reporting requirements that Western banks are. So you should be able to figure something out from the securities filings to HK regulators and other disclosure documents.

  • Posted by ZFC

    RH: “(a) a commercial exchange licence system (the banks would have to police that) ”

    Money changers exist and they are tied to commercial banks (they must be). And commercial bank reporting of FX to SAFE is extremely well administered.

    RH: “The obvious policy response could of course be that all fwd business would require 100% cash collateral in CNY”

    PBOC doesn’t care how banks hedge FX. Also, for the banks themselves, with regards to FX forwards, they do require order 10% collateral on FX forwards for their tier 2 customers. As you probably know, some customers require no collateral because they have big facilities to draw down on, some of which is apportioned to treasury products.

  • Posted by Twofish

    Huizer: I wonder if any of these aspects (1) deficient administration especially for credit risk management (2)weaknesses in design, enforcement and approval policy of FX control, have had attention in the Chinese language literature recently.

    It’s something that has been talked about, but there is an interesting dynamic in that because Chinese banks don’t have the risk control and management systems Western banks do, they are required to take fewer risks.

    One thing that has to be looked at is whether the risk and credit management systems that Western banks used actually reduced risk, or whether they led to a false sense of security and led Western banks to take risks that they should not have.

    One clear example of this is mortgage securitization where having a computer model telling you that everything is all right might prevent you from going out to talk to homeowners, look at all of the house flipping, and realize that everything isn’t all right and the computer models are giving bogus answers.

  • Posted by Pallj

    Twofish,
    you have a very valid point there. I only have first hand knowledge from the food industry, but there I have always said that your quality control systems are only useful if the people who operate them give a damn about what the product is like. If they don’t they’re soon just going through the motions.
    The banking sector may have been too plagued with key people caring more about the size of their bonus than the quality of their work. “If I don’t do it, somebody else will!” was their mantra for too many years, and we all know the result.

    The exponential growth of foreign assets makes me wonder how China has been able to pay for all of this. Relative to the size of China’s trade surplus the growth in foreign assets is so enormous it boggles my mind how all this foreign asset growth has been, and will be financed. A trillion would be four times their 2007 trade surplus, wouldn’t it?

Pingbacks