Since 2016, China’s financial authorities have been quietly pushing the banks to dispose of their bad loans more aggressively. The process has been incremental, but cumulatively meaningful. The measured pace at which Beijing is pursuing its debt cleanup could easily be misconstrued as timidity, a sign that China’s political leaders remain either uncommitted to dealing with the risk or unconvinced of the need. But the pace is deliberate. It is part of a broader, unconventional strategy designed to reduce NPLs in a way that minimizes both disruption to the economy and cost to the central government.
The approach is not without risk. While gradualism helps to avoid the trauma that would accompany a more sudden adjustment, the trade-off is that banks must tolerate a higher level of NPLs for a longer period of time, which affects the banking system’s liquidity. That also leaves the banks more vulnerable to economic shocks, which has the potential to lead to far greater economic disruption down the line. That said, Beijing has taken sweeping measures in order to mitigate those risks.
This report aims to make sense of China’s gradualist approach to cleaning up its financial system. In particular it will outline the ways in which Beijing’s approach is unconventional and what it’s trying to achieve. It will then discuss the steps the central government has taken to mitigate the sort of financial instability that could derail the process. And finally, it will look at how worsening economic conditions threaten the sustainability of the approach.
Dealing with bad loans is challenging for any government, and ignoring NPLs certainly has its advantages. That’s because write-downs of debt reduce GDP, diminish bank profits, and divert capital that might otherwise be used to expand lending. But failing to deal with bad loans in a timely manner can incur steep costs. Ultimately a day of reckoning comes—usually precipitated by some sort of financial crisis—when the accumulated weight of NPLs can no longer be ignored and the government must step in with a bailout.
China is pursuing a third way in addressing its bad loans. First, it has launched the debt cleanup prior to being forced into action by a financial crisis. By getting ahead of the problem, the authorities have been able to dictate the pace of the cleanup. Second, the Chinese government has avoided acknowledging the full extent of bank NPLs. That has frustrated observers trying to gauge the health of China’s financial system, but it’s part of an approach designed to minimize economic disruption. And finally, to avoid a government bailout—and more broadly to minimize the cost to the central government—Beijing has sought to pass on the cleanup cost either to the banks, the market or to local governments. I explain the last two aspects of this strategy in more detail below.
1. Minimizing Disruption – The Upside of Opacity
In the three years between 2016 and 2018, Chinese banks disposed of 4.4 trillion yuan worth of NPLs, equivalent to 4% of total outstanding loans at the end of 2018 (Figure 1). But it’s difficult to know how significant that 4% is as a portion of NPLs, since Beijing hasn’t disclosed the total volume of NPLs in the system. The official NPL ratio—1.83% at the end of 2018—is generally regarded as not reflective of reality. In recent years observers have offered estimates of NPL ratios that range from around 4% to as high as 20%. This wide range betrays just how little is known about the extent of the problem.
Figure 1. At Your Disposal
Annual NPLs in billions of yuan, and as percentage of outstanding loans at year end.
Source: CBIRC, Chinese media.
The official obfuscation serves a purpose. Back in 2013, I asked a senior financial regulator if the official NPL data fairly represented bank asset quality. At the time, the official NPL ratio of 1% seemed incredibly low given the malinvestment that resulted from the 4 trillion yuan stimulus following the global financial crisis. He didn’t answer my question directly, but instead asked, “Were China to disclose that it had significantly higher NPL levels, would anyone benefit?” He went on to explain that in that scenario, the banks would immediately have to raise new capital en masse and at firesale prices and companies would be forced into bankruptcy as banks pursued repayment. However, if the process could be drawn out by not disclosing the true state of NPLs, then much of the trauma could be avoided to buy authorities time to deal with the problem on their own terms and at their own pace.
No doubt he was somewhat overselling the potential trauma. Governments regulate their banks’ treatment of NPLs and hence can change capital and loan loss provision requirements to make it less burdensome for banks to recognize and hold bad loans on their books. However, China’s rules for dealing with bad loans are fairly strict. Banks are required to hold provisions equivalent to between 120% and 150% of expected losses on their bad loans (in contrast, European banks had a provisions coverage ratio of 46% at the end of June 2018 and an NPL ratio of 4.4%). In effect, China’s banks are forced to provision significantly more than they think they’re likely to lose on bad loans, which creates an incentive to dispose of NPLs quickly so they can reclaim the excess as profit.
However, such high provision requirements make it difficult for banks to recognize significantly higher levels of NPLs. Provisions come out of profits, but if profits aren’t sufficient then provisions reduce capital—hence the risk of equity firesales the regulator warned of. Of course, Beijing could just change the rules. But significantly weakening defences against bad loans potentially undermines trust in the system at precisely the wrong moment.
Instead, by maintaining that bad debt levels are significantly lower than is likely the case, regulators and banks can spread out the costs over time and deal with the problem at their own pace. It allows banks to raise fresh capital more slowly and with less urgency, allowing them to stagger their share offerings so as not to flood the market and drive down prices.
Similarly, not recognizing the bad loans all at once allows the banks to dispose of them over time, helping them to extract more value than would be possible if a sudden flood of bad loans were dumped on would-be investors all at once. It means that banks can gradually foreclose on delinquent borrowers and spread out bankruptcies—as well as the attendant job losses—over time. And it means that banks can use the profits they generate over a number of years to write down their bad loans, thereby reducing the need to raise capital.
2. Minimizing Costs – Passing the Buck
China’s central government has long been adept at devising creative ways to avoid using its own fiscal resources to pay for expensive policy measures. For instance, it relied on state banks, rather than its own coffers, to fund the 2008 stimulus. Moreover, it used foreign exchange reserves to recapitalize major banks after the last debt crisis. That creativity is on display once again in Beijing’s approach to NPLs as it shifts the costs of the cleanup.
The main cost has been borne by the banks, which have been able to pay for their write-downs of bad loans from profit, something that may not have been possible had banks been forced to recognize NPLs more quickly and in greater volumes. But private sector savings have also been enlisted in the NPL effort. Platforms such as Taobao, JD.com, as well as dozens of financial asset exchanges, have been auctioning bad loans directly to investors (see our reports here and here.) Securitization—a tool for selling NPLs to risk averse institutional investors—has also made a small contribution to NPL disposals, with 4% of NPLs disposed of in 2018 packaged into asset-backed securities (see our report here).
Perhaps the biggest new contributor to NPL disposals has been the creation of more than 50 local AMCs, bad banks that have the right to acquire NPLs from banks in bulk but are limited to operating in one province each. These provincial bad banks have been capitalized by a mix of local governments, state firms, financial institutions, and private companies (see here)—and notably not central government resources.
And to the extent that Beijing has been involved in the recapitalization process, it has done so in ways that have mobilized off-balance-sheet resources. In 2018, the National Social Security Fund acquired shares in both China Great Wall Asset Management Co. and China Orient Asset Management Co., two of the four original AMCs. National champions like China Telecom and China Property & Casualty Reinsurance were also mobilized as part of the recapitalization process.
While Beijing has so far dealt with NPLs relatively smoothly, its unconventional approach is not without risks. Disposing of NPLs slowly means that banks will likely have to maintain elevated levels of bad loans on their books for long periods of time. Under such circumstances, an economic shock could be potentially more destabilizing, particularly if it results in a sudden spike in NPLs or a sharp deterioration of liquidity conditions. Consequently, the authorities could be forced to abandon their slow, low-impact cleanup.
Beijing recognizes the risks, and has engaged in a broad-based effort to rein in financial risk in order to minimize the potential for crises. At the same time, it has augmented its liquidity management tools to ensure it can guarantee bank solvency.
1. Crisis Prevention
Improving financial stability has been a first order priority for Beijing since launching its deleveraging campaign in 2016. In fact, as we have written previously, the deleveraging campaign is best understood as a de-risking campaign—that is, a broad-based campaign to make China’s financial system safer, rather than an effort to reduce the overall debt burden.
Unquestionably, Beijing’s long-term goal is to reduce—or at least stabilize—debt levels relative to the size of the economy. However, given that China’s economic model is still dependent upon debt-enabled investment to deliver politically acceptable levels of growth, such a reduction is not likely to be feasible in the short term. Hence, the great success of financial reform has been—rather counterintuitively—to make the system less risky so that it might safely support higher levels of debt, at least for the time being (see Figure 2).
That’s not to say the deleveraging campaign has not involved actual deleveraging. Shadow banking has contracted, as has interbank lending, and NPL disposals are themselves a form of deleveraging. But rather than being a goal in itself, deleveraging has been deployed selectively as a tool to support the broader goal of making the financial system safer.
Figure 2. Fast Economic Growth Fueled By Even Faster Debt Growth
Change in total social financing from a year earlier, compared with change in nominal GDP
The China Banking and Insurance Regulatory Commission’s (CBIRC) de-risking effort has four prongs: reducing banks’ NPLs; mitigating the risk of financial contagion by rolling back lending between banks and other financial institutions; preventing the financial system from feeding credit into blatantly speculative activities; and migrating risk from shadow banks back to the formal banking system. All of this has massively reduced the amount of complexity in the system, put the regulators in a better position to manage risk, and generally diminished the likelihood of crisis.
2. Managing Liquidity
Financial crises are usually caused by a loss of confidence in financial institutions, not a surfeit of NPLs. However, NPLs can help drive that loss of confidence by creating liquidity problems for banks. Banks need the interest payments generated by loans to pay what they owe to depositors and to other funding sources. When loans go bad and those interest payments stop, banks might struggle to meet their obligations.
It’s not that the banks are broke. Rather, they likely have plenty of assets—namely, loans made to borrowers—but turning those assets into cash on short notice isn’t easy. Hence, banks might find themselves teetering on the edge of insolvency. In such a situation, the central bank can step in and lend to banks in return for collateral, ensuring that banks have sufficient cash to preserve confidence in the system.
China has greater control over market liquidity than other countries. Capital controls mean that the PBOC can print money and it won’t drain overseas. Meanwhile, printing money isn’t a politically charged issue in the way that it is in other countries, nor is government intervention in the markets. For instance, China’s authorities would not have hesitated to prop up Lehman Brothers. The consensus in Beijing is that stability comes before all else. That gives the PBOC the latitude to move more quickly than other central banks, should the situation demand its intervention.
The authorities can also impose solutions on market participants that might not be in their commercial interests. For example, Beijing can force distressed financial institutions into mergers on short notice, or it can demand that big banks keep lending to smaller financial institutions even in the midst of a cash crunch. To complement those advantages, the PBOC has in recent years built up an alphabet soup of lending tools (most notably the Standing Lending Facility and Short-Term Lending Operations) to ensure banks can borrow from the central bank whenever they need to.
Still, the provision of sufficient liquidity is merely a band-aid. It buys time for banks that must then use it to gradually reduce their accumulated NPLs. Despite Beijing’s best efforts, two related uncertainties still cloud the approach’s long term viability: the slowing economy and political resolve.
As growth slows, banks could find it more challenging to dispose of sufficient volumes of NPLs without government support or severe economic disruption. Meanwhile, senior leaders’ commitment to pursuing financial prudence over stimulating the economy could be seriously tested.
1. Whittling Down the NPL Stock
For Beijing’s gradualist approach to succeed, old NPLs must be disposed of at a faster rate than new ones are created. On one hand, that requires slowing the pace at which new NPLs are created, which has been a key aspect of the deleveraging campaign. But it’s somewhat difficult to gauge whether the pace of NPL accumulation is slowing because many NPLs are hidden or disguised as bank assets. As a group, these “assets”—a product of the shadow banking system—are probably of worse quality than loans, but they don’t show up in the NPL data. However, it’s likely the case that by reining in shadow banking, the CBIRC has helped curtail the generation of new bad loans.
On the other hand, banks must continue to sell, write-down, or otherwise dispose of their existing NPLs at a pace that allows them to reduce the overall burden, or else the whole effort may be futile. Of course, we have no way of knowing whether the banks have been able to realize that goal. Even if they have in the past, it’s likely to become more difficult as economic conditions deteriorate.
Slowing growth will likely result in more bad loans as companies that borrowed based on the expectation of continued fast growth will struggle to make loan repayments. Meanwhile, the plunge in the producer price index (PPI)—the measure of change in the prices of industrial and manufacturing inputs—over the past year has increased the real interest rate, therefore making it more difficult for many firms to service their loans (see Figure 3). And with the economy slowing, banks’ profit growth will likely slow as well. Replenishing loan loss provisions has already been eating up a progressively larger chunk of banks’ pre-tax profits in recent years (see Figure 4). For some smaller banks, profits might no longer be sufficient to sustain NPL disposals.
To be sure, banks are expecting NPL pressures to mount. Of those large banks that have published their 2018 earnings, most have disclosed significant increases in impairments against loan losses, in what effectively amounts to preparation for more write-offs to come. And some have been quite vocal in flagging the difficulties ahead.
“We deeply feel it’s quite difficult to maintain the low bad loan level. There are external factors, our own reasons, problems with multi layers of local governments and other pressure,” said Xu Yiming, chief financial officer of China Construction Bank, in March. “Do not think we are doing so well with 1.46 percent NPL ratio. It is very fragile. Once the environment changes, it can increase.”
Similarly, the market for buying NPLs might find itself overstretched as well. A bubble in NPL prices (which we wrote about here) popped in early 2018, and since then the pool of domestic NPL buyers has shrunk. Moreover, the local AMCs were relatively lightly capitalized when they were first established, and to maintain their rapid pace of NPL acquisitions they’ll likely need to raise more capital.
But the AMCs will likely face competition on raising capital, as analysts expect Chinese banks to raise large amounts of new capital over the next few years in order to maintain capital adequacy levels. That’s not primarily because of NPLs, but because the deleveraging campaign has required banks to rapidly expand lending to compensate for a contraction in shadow banking. This means a tighter capital raising environment, which could make it difficult for the AMCs and could conceivably result in the central government directly recapitalizing the bad banks.
Figure 3. Inflation Makes Debt Repayment Easier
Monthly change in China’s producer price index (PPI)
Figure 4. Bank Profits Have Been Sufficient To Meet NPL Write-Downs
Increase in loan loss provinces as a share of pre-tax profits at major banks.
Note: Banks must hold impairments on loan losses against NPLs. Impairments reduce profits. In the above chart, the smaller the percentage, the less impact impairments have on profit. A reading of 100% means that, were it not for impairments, pre-tax profit would be double.
2. Political Will
Since 2016, Xi Jinping has been willing to accept progressively slower growth in the interest of financial rectitude. The risk is that at some point growth will slow to a level where Xi—or the political consensus that exists around him—loses faith, and the deleveraging campaign is jettisoned. Some argue that Beijing has already pivoted back to stimulus mode with tax cuts, by front loading local government bond issuance into the start of the year, and by using window guidance to push banks into lending more to small firms.
It would be inaccurate to interpret an increase in the debt level as an abandonment of the deleveraging campaign. In fact, the very structure of the current stimulus speaks to Beijing’s commitment to cleaning up the financial system. Rather than a free-for-all, where banks and shadow banks are given the freedom to shovel as much credit as possible into the economy—which is broadly the approach to stimulus pursued repeatedly between 2009 and 2016—the current effort is targeted and limited only to banks (that have been chastened since their freewheeling days) and the bond market. The current effort to stimulate the economy has been designed specifically to avoid undoing the deleveraging of the last two years.
However, the targeted stimulus risks being insufficient in stabilizing growth. If the economy continues to suffer, Xi’s political resolve to hold the line on deleveraging could be sapped. If the political consensus around Xi evaporates, then there’s a risk that Beijing might again opt for old-style stimulus with the attendant increase in financial risk and bad investment decisions. Although there are no strong signs that this consensus is fraying, political considerations will nonetheless determine the sustainability of Beijing’s approach to the debt cleanup.
In the meantime, it seems likely that regulators will try to maintain their measured approach to dealing with NPLs and will continue to look for creative solutions to keep things moving. On one hand, they could try to make it easier for borrowers to service their loans. This is something recently flagged by Xu Zhong, director-general of the PBOC’s research bureau, when he suggested that lower interest rates were needed to help heavily indebted companies. He also suggested that local governments should be able to sell assets to raise the funds necessary to manage their debts. In fact, sales by private companies and state firms of both physical assets and equity in subsidiaries are something that could help with future debt repayments.
On the other hand, the government could make it easier for banks to dispose of their bad loans. It could approve local AMCs to launch IPOs so that they have the capital necessary to acquire greater volumes of NPLs. Or they could allow local AMCs to borrow directly from the interbank lending market, thereby reducing their borrowing costs. The Ministry of Finance could increase the volume of loans to the Big Four AMCs at discount interest rates, or even extend that privilege to some of the local AMCs. And it could even mobilize China Development Bank and the other policy banks to play a role in easing the stress on commercial banks.
Whether Beijing will use all the tools at its disposal is uncertain, but it clearly has options. Creativity has been the hallmark of the cleanup process thus far. Provincial AMCs, online NPL auctions, and credit card NPL securitization are all uniquely Chinese approaches to dealing with bad loans—and creativity is likely to remain Beijing’s greatest strength.
China is trying to avoid a state-led bailout of the financial system by cleaning up NPLs now, in a way that minimizes economic disruption and minimizes the burden on fiscal resources. So far it has proven to be moderately successful, and may end up being an inspired approach if it achieves the goal of gradually whittling down banks’ NPL burden.
However, dealing with NPLs represents only half the reform process. At the same time, China needs to change the way the economy grows so that it can deliver sufficiently fast growth with far less debt. That’s necessary for the long-term health of the financial system, but also to reduce the pain arising from the cleanup. In the meantime, Beijing will continue spreading the costs of dealing with NPLs around the economy—to the banks, private sector, local government, state firms, and other government entities—while striving to keep its own powder dry as the ultimate backstop.
 Michael Pettis, “What Is GDP in China?,” Carnegie Endowment for International Peace—China Financial Markets, January 16, 2019, https://carnegieendowment.org/chinafinancialmarkets/78138.
 Please note that the calculation is based on the CBIRC’s figure of 110.5 trillion yuan in outstanding loans at the end of 2018.
 China Orient Asset Management Corp publishes annual surveys of the NPL market in which it asks bankers what they believe is the actual level of NPLs in their area. In 2017 about 60% thought it was between 3% and 8%. See here for our report charting the survey’s results from 2014 to 2017.
 Ted Osborn, “Pressure on China’s banks to report bad debt is good news for foreign investors,” South China Morning Post, April 2, 2018,https://www.scmp.com/business/banking-finance/article/2139904/pressure-chinas-banks-report-bad-debt-good-news-foreign.
 “Supervisory Banking Statistics, Second Quarter 2018,” European Central Bank, October 2018, https://www.bankingsupervision.europa.eu/ecb/pub/pdf/ssm.supervisorybankingstatistics_second_quarter_2018_201810.en.pdf.
 Shu Zhang, Julie Zhu, “China’s top banks warn of ‘fragile’ bad loans scenario as economy slows,” Reuters, March 28, 2019,https://www.reuters.com/article/us-china-banks-results-idUSKCN1RA077.
 Liang Yu, “Deleveraging While Disseminating: The Task Facing China’s Banks,” S&P Ratings, October 18, 2018, https://www.spratings.com/documents/20184/0/RatingsDirect_DeleveragingWhileDisseminatingTheTaskFacingChinasBanks_39988427_Oct-19-2018.pdf/d644d494-c43f-29da-44ed-0c0661787142.
 Xu Zhong, “Xu Zhong: China Should Not Rush to Revamp Local Debt,” Caixin, January 22, 2019, https://www.caixinglobal.com/2019-01-22/opinion-china-should-not-rush-to-revamp-local-debt-101373076.html.
Get The Cleanup Report
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.