Bigger may not always mean better. The results of the 2019 stress test carried out by the People's Bank of China (PBOC) show that a few relatively bigger regional banks are under pressure, just like their smaller counterparts. The central bank's stress test covered 30 of the country's largest banks, each with assets of more than Chinese renminbi (RMB) 800 billion.

On analyzing the stress test results, S&P Global Ratings believes that some of the tested banks could require sizable recapitalization, just as was the case for Hengfeng Bank Co. Ltd. and Bank of Jinzhou Co. Ltd. In our view, Chinese banks are more vulnerable to idiosyncratic risks, arising from governance related issues and risk management deficiency, than systemic ones. Bank acceptance exposure and delayed financial reporting are among the key indicators of such risks. The quality of information and corporate action signals are important too.

Expanded Coverage Has Strengthened Monitoring

In our view, the PBOC has gained a better handle on China's financial risks, and 2019 was a year the regulator proactively experimented with bank bailout and recapitalization, rather than passively reacting to bank failure. The central bank's 2019 Financial Stability Report expanded the scope to 1,171 financial institutions (FIs) and performed single-factor sensitivity analysis on all these FIs, up from only 20 in 2018 (see table 1). Also, it increased the number of banks that were stress tested to 30 from 20. The 2019 stress test covered well over 80% of commercial banks' total assets.

Table 1

We believe the regulator's "severe" stress scenario for the 30 sizable banks it tested is informative and offers a reasonable level of stress. In more advanced markets, stress test scenarios typically assume recessions in their home markets, but these are mature economies growing at low single-digit rates. While a more severe stress scenario is possible in China should another crisis hit, the 4.15% GDP growth assumption that the PBOC tested represents a decline of 37% from 2018's GDP growth of 6.6%. That is much bigger than the 32% decline that China experienced during the global financial crisis of 2008.

Overall, we believe the PBOC adopted more realistic assumptions in the 2019 study than in 2018, such as assuming bigger increases in bank funding costs (see table 2). Though some assumptions could have been tougher still, such as currency depreciation, the central bank's severe stress scenario is a good starting point to examine potential risks among China's sizable banks, in our view.

Table 2

Tougher stress tests led to lower pass rates

More banks failed the severe scenario stress test in 2019 than in the previous year. That's because of the larger sample, which included more joint-stock and city commercial banks, tougher stress assumptions, and stricter nonperforming loan (NPL) recognition in 2019. The 2019 results show:

• Five banks would fail to meet the 5% common equity tier-1 (CET1) capital ratio requirement (compared to only one in 2018); and
• Seven would fail to meet the 6% Tier-1 capital ratio requirement (from two previously).

In addition, 10 of the 30 banks would also fail PBOC's liquidity test under the severe stress case. The test used more severe cash outflow assumptions for some off-balance sheet liabilities. It assumed drawdown factors that are 10 times those required under the liquidity coverage ratio framework. And, it tested funding window as short as seven days, which is inherently difficult for banks, given the intrinsic duration mismatch in their business model. This mismatch is why the liquidity coverage ratio framework has a longer funding window of 30 days. Accordingly, the PBOC said many of the FIs that failed the liquidity test have huge off-balance sheet exposure, and cautioned them to pay attention to their contingent financing obligations.

Though the banks that failed the stress test were not individually identified for fear of adverse market reactions, we assume Hengfeng and Jinzhou would have failed the test, given they were recapitalized last year after the test. We expect PBOC to require the other weaker banks that failed the test to shore up their capital.

Which Banks Could Follow Hengfeng And Jinzhou?

Several factors contribute to bank failures, especially those related to governance such as shareholder, related-party, or management issues, which are difficult to detect but could have serious consequences. The recapitalizations of Hengfeng and Jinzhou in 2019 offer insights into quantitative indicators that can signal which banks could be vulnerable. Jinzhou's case also provides figures with the belated filing of its 2018 and first-half 2019 financial reports to test these signals.

Bank acceptance exposure is an important quantitative risk indicator

In our view, high back acceptance (BA) exposure relative to total asset (TA) is a useful risk indicator. It could signal governance issues given the instrument's susceptibility to misuse, for instance as a tool for regulatory arbitrage. The BA-to-TA ratio at Hengfeng was 25% in 2016, while at Jinzhou it was about 15% in 2017 and 26% in 2018, before moderating slightly to 25% in first-half 2019 (see table 3).

Table 3  |  View Expanded Table

While BA is a legitimate form of funding to facilitate trades, this off-balance sheet commitment may be issued by banks (usually smaller ones) when they lack cash to make loans. The issuing bank gets some cash deposit via a margin guarantee without having the credit booked as a loan on its balance sheet, and the borrower can trade the paper for cash at a discount in the secondary market. Issuing BA is therefore akin to a bank printing its own "currency" (which is one of the reasons why a bank license is valuable to a local government). Meanwhile the instrument can be exploited by borrowers setting up shell trading companies to get funds.

We believe one of the reasons the PBOC took over Baoshang Bank last year was to draw the banking industry's attention to the risk of collateral damage stemming from smaller banks. A bank's BA has less "currency" across provinces as banks outside its province may not accept its paper (or do so at a huge discount). But the paper can still transmit the underlying bank's credit risk to the system, resulting in contagion through the interbank market. In July last year, the China Banking and Insurance Regulatory Commission (CBIRC) issued a notice alerting banks to the risks of shell intermediary companies and fictitious information in the bills business. In December, the regulator prohibited village banks from funding businesses or individuals outside the regions they are registered in, with the prohibition extending to lending and discounting commercial paper and BA.

We identify Bohai Bank, China Zheshang Bank, and Shengjing Bank as under more pressure than other PBOC stress-tested banks, given their significantly large BA-to-TA ratio. Location can amplify risks. For example, Shengjing Bank is based in Liaoning province and Bohai Bank in the mega-city of Tianjin, two regions where the local economy is under pressure. Bohai Bank and Zheshang are joint-stock banks while Shengjing is a city commercial bank.

Timely financial reporting matters

We consider corporate governance and transparency in the Chinese banking industry to be relatively weak. We reflect this risk in our rating anchor via the Banking Industry Country Risk Assessment (BICRA) score for China. In our opinion, delayed financial reporting signals potential management or shareholder issues. Hengfeng's last available financial report was for 2016. In the case of Jinzhou, its auditor Ernst & Young resigned because the bank failed to provide documents to confirm that its borrowers were able to service their loans amid indications of misuse of proceeds. And its 2018 financial report was released five months after the deadline.

When Jinzhou finally released its 2018 financial reports, they showed a swift and drastic deterioration in its financial condition (see table 4). Adjusting for special mention loans (SML), its loan loss reserves (LLR) covered only 28% of its problem loans (NPL plus SML) in 2018, down significantly from 84% in 2017. And it was much lower than an average of 104% for the stress-tested banks. Adjusting the bank's capital for LLR gap, defined as LLR less problem loans, would wipe out its CET1 capital, with its adjusted CET1 ratio at negative 1.78% versus the reported 6.07%. We believe this partly reflects the bank's extreme clean-up of its balance sheet (with NPLs exceeding all overdue loans, including those under 90 days) and the resultant provision charge.

Table 4  |  View Expanded Table

In Hengfeng's case, its LLR only covered 51% of its problem loans in 2016, and the bank's adjusted CET1 ratio would be 6.28% versus the reported 7.83%. Given the outdated financial statement, the gap is likely to be substantially bigger, as suggested by the recent RMB100 billion capital injection orchestrated by the government.

Looking beyond the numbers

In our opinion, adjusting a bank's capital position for the LLR gap helps to identify weak banks and anticipate their capital needs. At less than 7% adjusted CET1 ratio, Shengjing and Huaxia Bank seem more vulnerable to financial stress among the 30 banks tested. At an 8% threshold, the list expands to include Ping An Bank, Minsheng Bank, Guangfa Bank, Bohai Bank, and CITIC Bank, based on end-2018 figures.

All of these relatively vulnerable banks raised new capital in 2019, and the amount they raised closely matched their LLR gap (see chart 1).

Chart 1

However, this approach has its limitations, because it depends on the quality of financial information. The drastic change in the financials of Jinzhou attests to this. While Chinese regulators have tightened NPL recognition standards, banks still have wide discretion in defining other exposure, such as SMLs, for instance. Practices such as pooling and cross-guarantees also greatly increase a bank's risk, which reported numbers may understate.

In our view, the LLR gap is not a good gauge of capital needs for weaker banks with such risky practices. Jinzhou, for example, tapped the capital markets frequently, raising funds every year from 2015 to 2018.

Take the case of Zheshang, which reportedly built its off-balance sheet business by allowing borrowers to pool assets (usually BA) from related companies, mostly small private enterprises. The bank enabled borrowers to get funds quickly by tapping on the pool rather than applying individually, and cheaply through credit enhancement from cross guarantees. It has RMB250 billion of BA exposure off-balance sheet as of June 30, 2019, compared to only RMB12.6 billion raised from its A-share offering in November 2019. Thus, while Zheshang's capitalization might not seem vulnerable on paper, it may not be enough if bad loans go up, leading to additional write-offs that deplete its capital in the process of cleaning up its balance sheet. The bank's problem loan ratio rose to 3.34% in June 2019, from 2.83% in December 2018, highlighting asset quality deterioration.

Similarly, we believe more clean-up is in store for Shengjing, even though the bank seems to have recapitalized on paper. It received about RMB18 billion in fresh equity in November 2019 from a group of investors, including Evergrande Nan Chang, which received shareholder qualification approval from the Liaoning Office of CBIRC. We note that Shengjing's reported NPLs are only 39% of loans overdue more than 90 days in June 2019, and we expect the bank to accelerate its NPL recognition to meet regulatory requirements. It has RMB176 billion of BA exposure off-balance sheet as of June 30, 2019.

In the case of Bohai Bank, we find the quality of information to be lacking. Similar to other unlisted banks, Bohai Bank does not publish detailed semi-annual reports, and the brevity of its interim reports (spanning only six pages) belie the size of its assets, which exceed RMB1 trillion. The bank has yet to adopt international best practices with regard to financial reporting despite planning to list for many years and foreign bank ownership. Standard Chartered Bank owns about 20% of Bohai Bank since 2004, while local government-backed Tianjin TEDA Investment owns 25%. The last available figure of Bohai Bank's BA exposure off-balance sheet is RMB155 billion as of end-2018.

System Is Vulnerable To Slowdown, Contagion, And Idiosyncratic Risks

PBOC's 2019 stress test results confirm our view that China's NPLs are highly sensitive to GDP growth, even for its sizable banks. While the results provide only three data points (see chart 2), they illustrate the potential magnitude of the NPL challenge facing even its sizable banks should GDP fall to a 0%-2% range from 6.6% in 2018. China's GDP fell 4.5 percentage points during the global financial crisis. We estimate the existing problem loans at 6.5%-7.5% of the total.

Chart 2

Multi-factor hits to the system may undermine its resilience

PBOC's sensitivity analysis on all the 1,171 FIs included various factors, such as a general increase in NPL, a deterioration in asset quality in the property market or local government bonds specifically, or investment losses (see table 5).

Table 5

In our view, these factors are correlated. A broad-based decline in GDP driving up NPL is likely to hurt the property market as well as the finances of local governments, whose bonds account for about one-third of China's outstanding onshore bonds. This proportion is rising as a result of the refinancing of local government financing vehicle debt using local government bonds. The results of the sensitivity analysis are relatively benign for each factor (with capital adequacy above the 8% hurdle required per international norms). However, the broader financial system may be less resilient if we consider the contagion and indirect effects of the various factors.

We believe a severe scenario encompassing an eight-fold increase in overall NPL and a 15pp increase in local government bond NPAs may reduce the FIs' CAR by as much as 632 basis points (bps). The increase in overall NPL translates into an average NPL of 13.6% for the 1,171 FIs, from 1.7% at the end of 2018, while local government bond NPAs increase from a negligible level currently. Adding the 482bps and 150bps respective individual impact of these two factors would bring the FIs' average CAR to below 8% from a pre-stress level of 14.24% at the end of 2018. A severe deterioration in other factors concurrently as a result of contagion may further lower the CAR of these FIs, leading to more distress.

Idiosyncratic risks harder to spot but matter more

We have long since flagged risks associated with lending to complicated conglomerates, as reflected in our negative adjustment to the governance and transparency score in China's BICRA.

In our view, conglomerates' malfeasance presents idiosyncratic risks that are harder to spot. This concern is validated by the results of the concentration risk sensitivity, which the regulator added in its 2019 stress test to examine the impact of defaults by each FI's top borrowers.

The results show that the system is more vulnerable to concentration risk than local government bond or real estate risks in a severe stress scenario. Default by the top five borrowers in a severe stress scenario would lower the average CAR of the 1,171 FIs by 273bps, as compared to the 150bps impact from local government bond NPAs or 139bps impact from real estate NPLs.

New Regulation Will Rein In The Risks

We expect the PBOC to roll out new regulation to address the banking system's vulnerability to the governance, liquidity, and contagion risks revealed by its stress test.

In our opinion, weak governance is a common factor afflicting distressed banks, big and small. We found troubled banks Hengfeng and Jinzhou to have unusually high BA/TA exposure. This provides a good indicator for risks at all banks. The regulator's sensitivity test meanwhile demonstrates the system's vulnerability to concentration risk of large borrowers. We therefore anticipate more governance related rules in 2020, especially with regard to holding companies and conglomerates where transparency is low.

We expect more liquidity related rules to follow as well, particularly on liability management. Notwithstanding the relatively tough stress assumptions, having a third of China's sizable banks failing the liquidity test is likely to warrant risk mitigating measures from the regulator. Some of these measures may adopt a localized approach in their implementation to avoid unintended consequences that may destabilize the system. The notice alerting banks to the risks in the BA part of the business in July last year came from CBIRC's Beijing bureau, for instance.

The system remains vulnerable to contagion risk, and we expect the PBOC to implement as soon as practicable its plan to identify the domestic systemically important financial institutions in order to boost their capital buffers. Troubled banks will be offered liquidity support to repair their balance sheet, combined with management changes and curtailment of business growth until their problems are addressed, in our view. Local governments are often major shareholders of small banks, and the stronger governments would likely help bridge the capital gap. In addition to the local government's willingness to support, its ability to support is an important factor that determines the level of support and the type of support provided.

Distressed asset management companies have also been involved in providing support, as demonstrated in Jinzhou's recapitalization, and foreign capital is likely to be invited as well following China's Phase 1 trade deal with the United States. The regulator may allow a few banks to fail in cases where contagion risk can be ring-fenced, though such failures are unlikely to be abrupt given that deposit insurance is still at a nascent stage.

As China's economy slows down, a quick and proactive regulatory response will be critical to bolster the banking industry against severe stress.

APPENDIX

Table 6

Related Research

• Baoshang's Upcoming Tier-2 Debt Payment Will Shed Light On Banking Reform In China, Nov. 13, 2019
• Bank Of Jinzhou's Cancelled Preferred-Share Dividend A Wake-Up Call To Investors, Sept. 5, 2019
• The Coming Exit Of Struggling Banks, Aug. 27, 2019
• China Plays It Safe In Handling Troubled Bank Of Jinzhou, July 30, 2019