China's Foreign Exchange Reserves and Debt

One of my former colleagues dismissed Michael Pettis's work on the Chinese economy and international economics as "accounting focused" and therefore as too simplistic to be helpful in understanding the global economy. 

My view is that Pettis's focus on the key relationships between saving, investment and consumption across countries is a vital component of debate on the world economy because it shows the importance of imbalances - particularly between current account surplus and deficit countries - in causing problems in the global economy. 

I have referred to Pettis regularly on this blog and I admire his work because he highlights logical inconsistencies in some popular assumptions about the role of China and the United States and about the world economy generally.

One area that Pettis has highlighted (more than once) is the fallacy that China's reserves could be used domestically to pay off domestic debt. This fallacy has purchase because it sounds right. Why should the Chinese spend so much capital buying US Treasury securities, when they obviously have more important things to spend on within their own borders. This idea is particularly attractive at the moment with a deal not yet being finalised to ensure there isn't a default on US government bonds. This from Pettis's September 20th 2013 Newsletter.

China’s reserves only matter to its credit position if China faced a problem of external debt.

It doesn’t, and so the amount of reserves are almost wholly irrelevant, Because this argument seems to be reviving, it makes sense, I think, to repeat why central bank reserves cannot in any way help China resolve the crisis. I will leave aside the problems of whether the reserves are transferred in the form of foreign currency, in which case it does little to satisfy domestic RMB-denominated funding needs, or in RMB, in which case the PBoC must stop buying dollars in order to hold down the value of the RMB and in fact must sell dollars, which would cause the value of the RMB to soar, thereby wiping out the export sector in China.

A much more important objection is that the idea that reserves can be used to clean up the banks (or anything else, for that matter) is based on a misunderstanding about how the reserves were accumulated in the first place. There seems to be a still-widespread perception that PBoC reserves represent a hoard of unencumbered savings that the PBoC has somehow managed to collect.

But of course they are not. The PBoC has been forced to buy the reserves as a function of its intervention to manage the value of the RMB. And as they were forced to buy the reserves, the PBoC had to fund the purchases, which it did by borrowing RMB in the domestic market.

This means that the foreign currency reserves are simply the asset side of a balance sheet against which there are liabilities. What is more, remember that the RMB has appreciated by more than 30% since July, 2005, so that the value of the assets has dropped in RMB terms even as the value of the liabilities has remained the same, and this has been exacerbated by the lower interest rate the PBoC currently earns on its assets than the interest rate it pays on much of its liabilities.

In fact there have been rumors for years that the PBoC would technically be insolvent if its assets and liabilities were correctly marked, but whether or not this is true, any transfer of foreign currency reserves to bail out Chinese banks would simply represent a reduction of PBoC assets with no corresponding reduction in liabilities. The net liabilities of the PBoC, in other words, would rise by exactly the amount of the transfer. Because the liabilities of the PBoC are presumed to be the liabilities of the central government, the net effect of using the reserves to recapitalize the banks is identical to having the central government borrow money to recapitalize the banks.

This is the point. Any government that is able to borrow money can borrow money to recapitalize its banks, whether or not it has large amounts of foreign currency reserves. The amount of central bank reserves that China or any other country has is wholly irrelevant, except perhaps to the extent that without those reserves the central government would lack the credibility to borrow domestically, which hardly seems to be a concern in China’s case.

Bailing out the banks, it turns out, is conceptually no different than transferring debt from the banks to the central government. China can handle bad debts in the banking system, in other words, by transferring the net obligations from the banks to the central government, and the large hoard of reserves held by the PBoC does not make it any easier for China can resolve any future debt problems. In fact if anything it should remind us that when we are trying to calculate the total amount of debt the central government owes, the total should include any net liabilities of the PBoC, and that these net liabilities will increase by 1% of GDP every time the RMB strengthens against the dollar by 2%. ...    

So who is likely to cover the cost of NPLs in Chinese banks? This isn’t an easy question to answer. If the household sector continues to pay, either in the hidden form of repressed interest rates, or in the more explicit form of taxes, the existence of bad debt in the Chinese banking system must act to repress future household consumption growth. The transfers from the household sector to pay what may turn out to be a huge NPL bill will significantly lower the household income share of GDP, making it very unlikely that the household consumption share of GDP will rise.

If however the state sector covers the difference (perhaps by privatizing state assets and using the proceeds to pay down debt), we are left with the very difficult political problems, which China currently faces, of assigning the costs to different sectors or groups that control the state sector in China. The potentially very large cost of cleaning up NPLs must be assigned to groups that are likely to be both powerful and reluctant to pay the cost.

Debt always matters because it must always be paid for by someone – even if the borrower defaults, of course, the debt is simply “paid” by the lender. This is why the fact that debt in China seems to be growing much faster than debt-servicing capacity implies slower growth in the future. If the debt cannot be fully serviced by the increase in productivity created by the investment that the debt funded, unless it is funded by liquidating state sector assets it must cause a reduction in demand elsewhere, most probably in household consumption. This reduction in demand implies slower growth in the future and, of course, a more difficult rebalancing process.

Here Pettis is dealing with the argument that China's domestic debt problems can be easily solved. Another mistake is that China's foreign exchange reserves are historically unprecedented. This according to Pettis in a post from 2010 is also incorrect.

Twice before in history a country has, under similar circumstances, run up foreign reserves of the same magnitude.

The first time occurred in the late 1920s when, after a decade of record-beating trade and capital account surpluses, the United States had accumulated what John Maynard Keynes worriedly described as “all the bullion in the world”. At the time, total reserves accumulated by the US were more than 5-6% of global GDP. ...

The second time occurred in the late 1980s, when it was Japan’s turn to combine huge trade surpluses, along with more moderate surpluses on the capital account, to accumulate a stockpile of foreign reserves only a little less than the equivalent of 5-6% of global GDP. By the late 1980s, Japan’s accumulation of reserves drew the sort of same breathless description – much of it incorrect, of course – that China’s does today.

...  both previous cases turned out badly for long investors and brilliantly for anyone dumb enough to have gone short. During the early years of the Great Depression of the 1930s, US stock markets lost more than 80 per cent of their value, real estate prices collapsed, and the US economy contracted in real terms by an astonishing 30-40 per cent before recovering in the 1940s.

Japan’s subsequent experience was economically less violent in the short term, but even costlier over the long term. During the period following its astonishing accumulation of central bank reserves, its stock market also lost more than 80 per cent of its value, real estate prices collapsed, and economic growth was virtually non-existent for two decades.

The idea that massive levels of reserves are a guarantor of economic stability is, in other words, based on a profound misunderstanding both of history and of the nature of reserves. Reserves of course are not useless as an enhancer of financial stability, but their use is for very specific forms of instability. Having large amounts of reserves relative to external claims protects countries from external debt crises and from currency crises.

Great, but neither Chanos, nor even the most pessimistic Sino-analyst, has ever said that these are the kinds of risks China faces today, any more than they were the risks faced by the US in the late 1920s or Japan in the late 1980s. The risks that China faces today (and the US in the late 1920s and Japan in the late 1980s) is of excessive domestic liquidity having fueled asset and capacity bubbles, the latter requiring the uninterrupted ability of foreign countries to absorb via large and growing trade deficits. These risks include an explosion in domestic government debt directly and contingently through the banking system.

These are, very typically, the kinds of risks that threaten rapidly developing large economies, unlike the external debt and currency risks that typically threaten small economies. And reserves are almost totally useless in protecting these economies from the risks they face (and, no, no, no, reserves cannot be used to recapitalize the banks – only domestic government borrowing or direct or hidden taxes on the household sector can be used to recapitalize the banks).

In fact, it was the very process of generating massive reserves that created the risks which subsequently devastated the US and Japan. Both countries had accumulated reserves over a decade during which they experienced sharply undervalued currencies, rapid urbanization, and rapid growth in worker productivity (sound familiar?). These three factors led to large and rising trade surpluses which, when combined with capital inflows seeking advantage of the rapid economic growth, forced a too-quick expansion of domestic money and credit.

It was this money and credit expansion that created the excess capacity that ultimately led to the lost decades for the US and Japan. High reserves in both cases were symptoms of terrible underlying imbalances, and they were consequently useless in protecting those countries from the risks those imbalances posed.

We must be careful how we read history. The fact that the US and Japan had terrible decades following periods during which they had amassed levels of reserves that China has subsequently matched, and under conditions similar to those of China, does not necessarily mean that China too must have a lost decade or two. ... 

But the history does indicate that facile statements about central bank reserves should, at the very least, be measured against the obvious historical precedents.

In an earlier post Pettis explains in more specific details the process of foreign exchange reserve accumulation. In a 2012 post entitled "What the PBoC Cannot Do with its Reserves", Pettis starts by explaining how the Peoples Bank of China (PBoC - China's Reserve bank) goes about keeping the renminbi (RMB) (the yuan if you prefer - see here for explanation) at a lower level.

As long as China ran the largest current account surplus ever recorded as a share of global GDP, and the US the largest current account deficit ever recorded, and especially since China also ran an additional capital account surplus (i.e. other non-PBoC agents ran a net capital inflow), it was almost impossible for the PBoC to do anything but buy US dollar assets. Given the sheer amounts, a substantial portion of these assets had inevitably to be USG bonds. 

This was not a discretionary lending decision. It is the automatic consequence of China’s currency regime, in which it pegs the RMB to a foreign currency, in this case the dollar. Why? Because when the PBoC decides on the level of the RMB against the dollar, it does not do so by passing a law, and making it a capital crime for anyone to trade at a different price. What it does is far simpler. It offers to buy or sell unlimited amounts of RMB against the dollar at the desired price. 

No one will sell dollars for less than what they can get from the PBoC, nor will anyone buy dollars for more than what they can pay the PBoC, so all transactions get done at that price. That is how the PBoC (or any other central bank that intervenes in the currency market) sets the foreign exchange value of its own currency.

Now the reason the PBoC must buy dollars under the regime is because the rest of China is a net buyer of dollars. Remember China runs a current account surplus, which must be matched by a net capital outflow.

This means that as long as it wants to set the exchange rate, then, it must take the opposite position of the market. Since the rest of the market is a net seller of dollars (China runs a current and capital account surplus), the PBoC has no choice but to be a net buyer of dollars, which of course it must then invest. 

If it stops buying dollars, it must let the market decide by itself on the new equilibrium price of the dollar. In that case the value of the dollar has to plunge in RMB terms (or the RMB soar, which is the same thing) in order for buyers and sellers to match up and for the market to clear. The moment the PBoC stops buying, in other words, the RMB will rise in value – and so it cannot stop buying in anticipation of the RMB rising in value.

 The next issue is how the PBoC funds these purchases of dollars:

It does so primarily by borrowing in the domestic money markets, selling PBoC bills or entering into short term repos (although it also issues some longer-term bonds), or by “creating” money by crediting the accounts of the commercial banks who sell it the dollars. 

This means, to simplify, that the PBoC has a balance sheet consisting on one side of dollar assets (and here “dollar” is short-hand for all foreign assets). Against this and on the other side it has a roughly equivalent amount of RMB liabilities (I say “roughly” because when you run a mismatched balance sheet, changes in the relative value of assets and liabilities will create losses or profits). 

Here is where things get interesting. China’s reserves are often thought of as if they were a treasure trove available for spending. They are not. They are simply the asset side of the mismatched balance sheet. If the PBoC wanted to “spend” $100, say for example to recapitalize a bank, it could do so, but this would automatically create a $100 dollar hole in its balance sheet. – it would still owe the RMB that it borrowed originally to purchase the $100. To put it another way, the reserves are not a savings account, free for the PBoC to spend as it likes. Reserves are effectively borrowed money. ... 

So what are reserves good for? As long as China maintains its own currency and denominates all domestic transactions in RMB, the PBoC reserves cannot be used in China. They cannot go to pay doctors’ salaries, to build bridges, to lower taxes or to subsidize consumption. They can only be used to purchase or pay for things from outside China. This means that reserves ensure that China can import foreign commodities and other goods as long as it can pay for them domestically. It also means that the PBoC can ensure the availability of dollars to repay foreign debt and foreign investment. ... 

Reserves are useless in preventing domestic debt crises (not totally, because they affect the credibility of the currency, but the RMB today doesn’t seem to suffer from a lack of credibility).

Pettis then goes on to explain why a revaluation of the renminbi against the dollar does not cause the simple scenario of huge losses for China and huge gains for the United States that most commentators allude to when they discuss the so-called economic balance of terror between the United States and China. There are still winners and losers, but the equations are more complex than most assume.

Many people in China and abroad have argued that China cannot afford to raise the value of the RMB against the dollar because it would mean that China will take huge losses because of its massive reserves. After all, if the RMB rises by 10% against the dollar, the value of its reserves will have necessarily declined by $250 billion in RMB terms.   

This is almost completely wrong – China will not take losses anywhere close to that amount and may probably even take a gain if it revalues the currency. One foreign economist even published a rather loony piece three months ago, which excoriated the Obama administration’s “bogus” trade argument for revaluation as done purely for nefarious and no doubt imperialistic reasons – and to strengthen the conspiratorial air it somehow ignored the fact that nearly every country in Europe and Asia has made the same argument.

The argument is that the US wants China to revalue the RMB not because of trade rebalancing (wrong, and this makes a common but still annoying mistake about the relationship between the currency and the trade balance) but rather because of a secret American scheme to reduce the amount that the US government has to pay China on its PBoC holdings. Appreciation of the RMB, according to this theory, represents a transfer of wealth from China to the US because it effectively reduces cost to the US of servicing the debt

An appreciation of the RMB cannot reduce the cost of the US government's debt obligations because:

The US government transacts almost exclusively in dollars, raises dollars in the form of taxes and borrowing, and owns dollar assets. Since it will pay exactly the same number of dollars to Chinese investors after the change in the RMB value as it did before the change, simple arithmetic should indicate that there will be no impact at all on the cost to the US of repaying the debt. But this doesn't mean there aren't winners and losers including within China. Working them out requires an analysis of the "various balance sheets". 

So who are the winners and losers?

In a nutshell, anyone who is net long dollars against RMB loses, and anyone who is net short dollars against RMB gains. For China, the equation is the same. Those who are short US dollars will gain. 

There is no precise way of answering this question, because every single economic entity in China implicitly has some complex exposure to the dollar (by which I mean foreign currencies generally) through current and future transactions, but generally speaking China is likely to gain from a revaluation because after the revaluation it will be exchanging the stuff it makes for stuff it buys from abroad at a better ratio. The value of what it sells abroad will rise relative to the value of what it buys from abroad, and if we could correctly capitalize those values on the balance sheet, it would probably show that the Chinese balance sheet would improve with a revaluation of the RMB. 

Some people might make a more sophisticated argument that since China is a net creditor – i.e. it is net long dollars – it will lose by a revaluation of the RMB. This argument also turns out to be wrong, but for more complex reasons, and to explain why I have to put on my former-trader’s hat and explain the difference between a real loss and a realized loss.

 This is where it gets even more complicated.

If you believe that the RMB is undervalued then you must accept that China takes a “real” loss every single time it exchanges a locally produced good or asset for a foreign one. It does not “realize” the loss, however, until it revalues the RMB to its “correct” value. 

In other words, the PBoC, as the representative of China’s net creditor status, will immediately realize a loss when the RMB revalues, but this loss did not occur because of the revaluation. It occurred the very day the trade took place. When a Chinese producer sold goods to the US and took payment in US dollars, there was an unrealized economic loss equal to the undervaluation of the RMB. This unrealized loss was passed onto the PBoC when it bought the dollars from the exporter and paid RMB. 

This loss, however, will not actually show up until the RMB is revalued, which forces the real loss to be realized (i.e. recognized as an accounting matter). Postponing the revaluation, then, is not the way to avoid the loss – it is too late for that. The only way to avoid future additional loss is to stop making the exchange, which means, ironically, that the longer the PBoC postpones the revaluation of the RMB, the greater the real loss it will take. 

So a revaluation of the RMB will not cause any real loss to any Chinese entity today. The loss already occurred but hasn’t been realized. 

But wait, if the RMB is revalued by 10%, the value of the PBoC’s assets will immediately decline by $250 billion in RMB terms. Since the Chinese measure their wealth in RMB, isn’t this a real additional loss for China? 

No, because remember that the only thing you can do with reserves is pay for foreign imports or repay foreign obligations. And just as the value of the reserves drops 10% in RMB terms, so does the value of all those foreign payments – by definition they must go down by exactly the same amount in RMB terms. 

This means that China takes no loss. It can buy and pay for just as much “stuff” after the revaluation, and with less implied PBoC borrowing, as it could before the revaluation – and the real value of money is what you can buy with it. So the real value of the reserves hasn’t changed at all – just the accounting value in RMB, but this simply recognizes losses that were already taken long ago when the trade was first made, and should be a largely irrelevant number (except perhaps for conspiracy theorists). 

But there are important impacts within China. Who wins and who loses depends "on the structure of individual balance sheets."

Basically everyone who is net long dollars against the RMB loses in an appreciation, and everyone who is net short dollars against the RMB wins. 

Who loses? Of course the PBoC is a big loser. ...

Exporters and their employees, too, are naturally long dollars and so they would lose. ...

Chinese companies with more assets abroad then foreign debt might also lose. 

Who wins? Nearly everyone else in China, since everyone in the country is short dollars to the extent that there are imported goods in his life. The local tea seller is short dollars if his tea is delivered to him in gas-guzzling trucks, as is the family planning to visit Egypt next year, as is the local provider of French perfumes, as is a teenager who wants to buy Nike shoes, and so pay for the corporate sponsorship of a Brazilian soccer star playing for a Spanish team. Every household and nearly every business in China is, in one way or another, an importer (and this is true in every country), so unless they own a lot of assets abroad they are effectively short dollars and will benefit from an appreciation in the RMB. 

Revaluing the RMB, in other words, is important and significant because it represents a shift of wealth largely from the PBoC, exporters, and Chinese residents who have stashed away a lot of wealth in a foreign bank, in favor of the rest of the country. Since much of this shift of wealth benefits households at the expense of the state and manufacturers, one of the automatic consequence of a revaluation will be an increase in household wealth and, with it, household consumption. This is why revaluation is part of the rebalancing strategy – it shifts income to households and so increases household consumption.

So a revaluation has important balance sheet impacts on entities within China, and to a much lesser extent, on some entities outside China. But the fact that the PBoC loses big time does actually matter and although Pettis doesn't mention it, it would actually lead to the same sort of structural changes in the Australian economy that many people in Australia are worried about right now - namely a decline in manufacturing and (traded) services competitiveness. Pettis's point probably is that the rise in household wealth would help to balance the economy away from the investment and export dominated growth model.

But since it merely represents a distribution of wealth within China should we care about the PBoC losses or can we ignore them? 

Unfortunately we cannot ignore them and might have to worry about the PBoC losses because, once again, of balance sheet impacts. 

The PBoC runs a mismatched balance sheet, and as a consequence every 10% revaluation in the RMB will cause the PBoC’s net indebtedness to rise by about 7-8% of GDP. This ultimately becomes an increase in total government debt, and of course the more dollars the PBoC accumulates, the greater this loss. (Some readers will note that if government debt levels are already too high, an increase in government debt will sharply increase future government claims on household income, thus reducing the future rebalancing impact of a revaluation, and they are right, which indicates how complex and difficult rebalancing might be). In that sense it is not whether or not China as a whole loses or gains from a revaluation that can be measured by looking at the reserves, and I would argue that it gains, but how the losses are distributed and what further balance sheet impacts that might have.

Simple right? If you're confused, you're not alone. It is easy to see why so many writers assume simple effects and consequences of actions because the real world complexity of a change in the value of the RMB is much harder to explain. Much easier just to say China loses from a revaluation of the yuan against the dollar!

Still I wouldn't want to betting on the certainty of anything in this field. As Keynes once said: "Markets can remain irrational a lot longer than you and I can remain solvent."