One key question around China is pretty straight forward: will losses in China’s banks and shadow banks—whether on their lending to Chinese firms or their lending to investment vehicles of local governments* necessarily give rise to a currency crisis? Or can China, in some sense, experience a banking crisis—or at least foot the bill for legacy bad loans—without a further slide in the yuan (whether against the dollar or against a basket)?
To answer this question I think it helps to review the reasons why banking crises and currency crises can be correlated, and to see what vulnerabilities are and are not present in China.
The first reason why a banking crisis can lead to a currency crisis is simple: the banks have financed their lending boom by borrowing from the rest of the world, and the rest of the world decides the banks are too risky and wants its money back. The need to repay external creditors leads the country to exhaust its foreign exchange reserves, and ultimately, without reserves, the country is forced to devalue. Thailand in 1997 is probably the best example.
This risk simply is not present in China. China has more external reserves than it has external debt, let alone short-term external debt. China's lending boom hasn’t been financed by the world—it has been financed out of China’s own savings, intermediated through Chinese financial institutions.
The second reason is also straightforward: losses in the banks and shadow banks could lead Chinese residents to pull their funds out of China’s financial system, and seek safety offshore.
This no doubt could happen—though China’s financial controls are meant to limit this risk. China—like other big countries—doesn’t have enough reserves on hand to cover all its domestic bank deposits, let alone the shadow banking system’s analogue to “deposits.” And while some deposit flight can be financed out of China’s existing trade surplus, it is certainly possible to imagine more flight than could be financed out of China's exports.
On the other hand, losses in China’s domestic banking system will not necessarily result in a run into offshore deposits. The system may be recapitalized before there is a run. Or those who flee the shadow banking system might move their funds into China’s banks, not into foreign deposits. Or those who flee China’s risker mid-tier banks might run to the safety of the big state commercial banks (effectively running out of institutions backed by weak provincial government balance sheets to institutions backed by the much stronger balance sheet of the central government).
But a run out of all Chinese bank deposits is a risk, both to China and the world. It is in some sense is the flip side of China’s lack of external vulnerability: very high domestic savings intermediated through domestic financial institutions means a ton of domestic deposits and shadow deposits. And limiting this risk is a big reason why I believe China needs to be cautious in liberalizing its financial account.
The third reason is that China’s government might not be able to cover the cost of recapitalizing its banks, and the government—not the banks per se—might need to turn to the central bank for financing. This is one of the risks that Christopher Balding highlights for example (more here). And while it is a risk, I don’t think it is a big risk.
A bank doesn’t actually have to be recapitalized with cash. It can be recapitalized with government bonds (see Jan Musschoot for the mechanics, or look at this IMF paper). Say a bank writes down the value of its existing loans, and that loss wipes out its equity capital. The government can exchange government bonds for “new” equity in the bank. It doesn’t have to go out into the market and sell bonds and hand the cash over to the bank.
An asset management company can also be funded in the same way: the government can swap newly issued government bonds directly for a portfolio of bad loans (and hand the bad loans over to an asset management company to try to recover something). This raises the government’s stock of debt, but it doesn’t require raising cash and handing the cash over to the bank in exchange for a portfolio of bad loans. It also doesn't require making use of the central bank's balance sheet.**
And even if the government wants to recapitalize its banks by handing the banks cash in exchange for either new equity or for bad debts, it can raise the cash by issuing bonds in the market—that doesn’t require a monetary expansion either, though it can put upward pressure on interest rates. The IMF's 2016 estimate of bank losses on corporate credit (7 percent of China's GDP) may be too low, but if it is close to right, it is not a sum that China would have trouble funding.***
To be clear, if a recapitalized bank experiences a run, the bank will need to take the government bonds it has received from the government to the central bank and borrow cash against its “good” collateral (or not-so-good collateral; I agree with Balding's World that a no-recourse loan against bad collateral is a backdoor bank recapitalization through the central bank). But it is the run that gives rise to the need “to print” money, not the recapitalization. And the money provided to depositors fleeing a troubled institution often ends up in other institutions—it doesn’t necessarily leave the system. The central bank can mop up liquidity provided to a troubled institution by withdrawing liquidity elsewhere, with no change in its monetary policy stance.
One additional point here: a preemptive recapitalization which adds to the system’s capital and allows some shadow banking liabilities to migrate on-balance sheet would in my view reduce the risk of a run—as it would be clear that the recapitalized institutions would be able to absorb losses without passing the losses on to depositors. It thus in my view reduces the risk that the banking system's legacy bad loans would lead to a monetary expansion that jeopardizes currency stability.
The fourth reason why a banking crisis can lead to a currency depreciation is that the banking crisis leads to a slowdown in growth—and in response to the slowdown in growth, the central bank may need to ease monetary policy. Capital controls can give a country with a currency peg a bit more space to keep its currency stable without following the monetary policy of its anchor currency (or for a basket its anchor currencies). For example, for much of the last 15 years, China has been able to have a tighter monetary policy—or at least higher lending rates—than the United States without being overwhelmed by inflows (from 2003 to 2013, China’s challenge was limiting inflows, not outflows). But there is a limit to how much any country, even China, can ease monetary policy while maintaining a stable exchange rate, especially if China is managing its currency against the dollar, and the U.S. is tightening monetary policy. China’s controls can make it significantly harder to swap yuan for dollars or euros, but they are likely to work best if the controls are reinforced by a positive interest rate differential.
Here too China has options. It could respond to a slowdown in growth by easing fiscal policy without easing monetary policy, maintaining an interest rate differential that would encourage Chinese residents to keep their funds in China.*** And that could maintain demand—taking pressure off the central bank. In any case, the PBOC is now tightening monetary policy to slow the economy, so this is a theoretic rather than a current risk.****
While there is a path out of China’s current banking troubles that doesn’t involve a further depreciation, there isn’t a path out of China’s current difficulties that doesn’t involve the use of the central government’s balance sheet. *****
Let me offer up an imperfect analogy—imperfect both because it involves a currency union that isn’t a full political union, and even more imperfect because it involves a currency that floats, not a peg. Ignore it if you want, my argument doesn’t depend on it.
Before its crisis the eurozone ran a balanced current account. The current account deficits of countries like Greece, Ireland, and Spain were essentially financed (in euros) by German and Dutch current account surpluses, not by borrowing from the rest of the world. And the run out of Greek, Irish, and Spanish banks in 2010 and 2011 was largely a run into safe assets in the eurozone’s core, not a run out of the euro. That all was a big reason why the euro didn’t depreciate significantly in the early phases of the eurozone’s crisis, despite violent swings in financial flows inside the eurozone. Keeping the eurozone together required the ECB act as a lender of last resort (essentially borrowing from German banks to lend to Spanish and Italian banks through the target 2 system to offset the withdrawal of private financing from the periphery) and that the eurozone create common institutions (EFSF, ESM) to help weaker countries finance the cost of bank recapitalization. But the ECB’s provision of lender of last resort financing to banks in troubled countries on its own did not drive the euro down.
The euro ultimately did fall in 2014 because the ECB needed a looser monetary policy to support overall eurozone demand (negative rates, QE, etc). If the eurozone as a whole had relied more on fiscal rather than monetary easing to rebuild demand, the ECB wouldn’t have needed to ease quite as much—and the eurozone today would have a smaller current account surplus.
I think there is a parallel: China’s shadow banks and some mid-tier banks are the periphery, relying on funding from China’s core (so to speak). A run back to the core is no doubt a significant problem. But it also is something that conceptually China has the resources to manage without necessarily needing a weaker currency and more support for its growth from net exports.
* China’s central government's credit risk is low; central government debt is low—and lending to Chinese households also isn’t generally believed to pose a problem.
** The asset management companies (AMCs) that were set up to clean up the balance sheets of the major state commercial banks initially had this structure: the banks handed over their bad loans to the AMCs, and got a bond that the AMCs issued in exchange. The AMC bond was never explicitly guaranteed, so technically it wasn’t the government’s debt. But the government pretty clearly was going to stand behind the AMC loans. There was no direct need to use the PBOC’s balance sheet in this transaction. Christopher Balding notes that the central bank can also provide liquidity directly to the banks against dodgy collateral, and thus lift bad loans directly off a troubled bank's balance sheet (either by buying the bad loan, or by providing a no-recourse loan against the loan). That is no doubt true: China has been known to hide the cost of a bailout by in effect netting it against the central bank's ongoing profits in a less than transparent way. But the orthodox way of structuring an AMC would use the Ministry of Finance's balance sheet, and the central bank would lend against recapitalization bonds or AMC bonds with a guarantee not directly against bad collateral. And any injection of liquidity to a troubled bank would be offset by withdrawing liquidity elsewhere. For those interested in the details of China's recapitalization of the big state banks, there is no better source than Red Capitalism.
*** China is now big enough that a slowdown in its growth affects growth elsewhere, and thus monetary easing by China's partners also might play a role in maintaining the interest rate differential. In 2016 for example, risks around China seem to have contributed to the Fed's decision to slow its pace of tightening.
**** A couple of additional technical points here. In 2015 and in early 2016, the PBOC was loosening policy not tightening policy (cutting rates, reducing the reserve requirement and so on). That added to the pressure on China's currency. And with reserves falling, the PBOC needed to buy domestic assets (or increase its domestic lending) to keep its balance sheet from shrinking. Its overall monetary policy stance consequently cannot be inferred by looking only at its domestic balance sheet.
***** A restructuring of local government debt also does not require the use of the central bank's balance sheet. For example the central government could swap a Ministry of Finance bond for provincial debt, leaving the market (read banks and shadow banks) with a claim on the central government and leaving it to the central government to collect on provincial debt.