- Expect government’s intervention and banks’ provisions to absorb the potential loss without hurting earnings in base-case scenario
- Expect c.9.5% earnings impact in an unlikely worst-case scenario that banks afford all
- Near-term volatility to remain on slower property market recovery
- Maintain positive outlook in the long-run for China banks, especially retail banks like CMB (3968 HK) and PSBC (1658 HK)
Systematic risk unlikely and earnings impact not as big as thought.
The banking sector was hit by the mortgage suspension news. Our analysis shows that the value of mortgages impacted might be as high as Rmb 1.1tr, while the banking system could absorb Rmb 7.5tr before triggering a systematic risk. We believe it’s highly likely that the government will intervene and Rmb 1.1tr monetary support will be needed. In our base case that banks will share 30% of the monetary support over 3 years, banks’ provisions could well absorb the potential impact without hurting earnings. In the worst case scenario where banks bear all the bad debt, the earnings impact would be c.9.5% in FY22F.
Most banks would maintain solid capital ratio under the stressed scenario.
Though the market has been concerned on CMB (3968 HK) and PSBC (1658 HK) as they have the highest mortgage exposure over total loans, the impact will not be higher on these 2 banks given 1) their exposure to developers that have defaulted and the central region is not that high; and 2) they have very high provision coverage ratio of >400% to cover potential losses. The Big Four banks have strong capital ratios and could well maintain their dividend payouts. Among our coverage, CITIC (998 HK) and CEB (6818 HK) are relatively weaker and may not be able to maintain their dividend payouts in the worst case scenario.
Maintain positive outlook in the long run but prefer HK banks at this juncture.
Even if there is limited impact on earnings, we expect the sector to face near-term volatility as market may continue to worry about future mortgage growth and further delays to the property market recovery. We turn more neutral on China banks and prefer HK banks. Still, longer term, we remain positive on China’s economic recovery and banks’ performance, especially retail banks like CMB and PSBC, which are trading around 5-year lows.
How much buffer in the banking system?
Recent events raised investors’ concerns on banks’ risk.
Recent news of 1) village banks’ scandal in Henan and 2) mortgage repayment suspension by homebuyers of some undelivered units, triggering concerns on banks’ credit risk. The China banking sector saw high price volatility in the past week led by mortgage repayment suspension that was spreading quickly across the country. Currently, we do not see systematic risk materialising as the buffer in the banking system is high enough to cover potential losses.
Banking system can absorb at least c.Rmb7.5tr before triggering a systematic risk.
In our base case of “systematic risk”, we expect CET 1 ratio (Common Equity Tier 1/ Risk weighted assets) to be 7.5% and the loan loss coverage ratio to be 150%. As of end of 1Q22, the CET 1 ratio in China’s banking system was 10.7% and coverage ratio was 200.7%. With high capital ratio and impairment coverage ratio, we expect the system to be able to absorb c.Rmb 7.5tr of loss triggering systematic risk (Fig 1).
In fact, we are taking a conservative view to define the systematic risk level. The lowest requirement for CET 1 ratio is 5% while we already have a 2.5% buffer. The lowest coverage requirement is 120-150% and we are also using the higher end for our calculation. If we assume a lower requirement (5% CET 1 ratio and 120% coverage ratio) to assess systematic risk level, the loss absorbing capacity will be Rmb 13tr. And this has not factored the banks’ net earnings for this year, which will also be used to absorb losses before the erosion of banks’ capital.
How large is affected mortgage exposure?
Very small exposure to affected projects as reported by banks.
Most banks have made announcements on Jul 14 that their exposure to affected projects is very small. Most reported <0.1% exposure (Fig 2).
However, given the number of projects involved in mortgage payment has grown quickly and no longer traceable, investors may be concerned on what is the exposure on a worst case scenario if more projects are involved. We have done a stress test to find out the maximum loss banks may have to bear though the actual probability for the worstcase assumptions to take place is very low.
Banks may have c.Rmb1.1tr of mortgages under delivery risk in base case scenario of the stress test.
Our base case assumption is 30% of projects presold by liquidity-stretched developers are subject to delivery risk. We expect this will mainly impact new mortgages issued after 2020, which was c.Rmb 8.77tr, or 23% of c.Rmb 38.8tr of outstanding mortgage balance as of the end of 1Q22. The mortgage
value subject to delivery risk would be c.Rmb 1.1tr or 2.9% of existing mortgages.
Please note that this is a just a rough estimate on the banking system’s overall exposure and we have not
considered the major banks’ preference and exposure to Tier 1/2 cities and quality property developers. This also ignores that the government could step in to ensure delivery, which is a scenario overlooked by many investors.
Will government step in?
Highly likely the government will step in.
In fact, we should not look at the worst-case scenario without considering that the government could step in to ensure delivery. The key request from home buyers is the delivery of their flats. Threats of suspension of monetary payment is just a way of being heard. Banks and home buyers have the same goal on this. The China Banking and Insurance Regulatory Commission (CBIRC) has already emphasized that it will guide financial institutions to work together with the government to solve the problem and “ensure delivery” as the first priority.
We expect the government to step in to solve this problem, as it may affect >Rmb 3tr of residential sales value and c.4m families. The government will have to stabilize the situation before things this builds up to an uncontrollable scale. Stabilizing such events has always been a priority of the Chinese government. The village banks’ scandal in Henan was brought under control by the government’s announcement that affected individuals will be repaid.
Expect Rmb 1.1tr to contain delivery risk.
According to DBS property team’s analysis (link), in the base case scenario (30% of project presold by liquidity strapped developers are subject to delivery risk and 30% of construction costs have been settled upfront), the monetary support required may amount to Rmb 1.1tr. Considering that the fiscal income in
5M22 was Rmb 8.67tr, government can afford this amount, especially when 1) the amount is likely to be spread over a few years; and 2) there are other parties like banks and developers to share the costs.
What are the likely impacts on banks?
Base case scenario assumption: Government intervenes, banks take 30% of monetary support, spread over 3 years.
Our base case scenario analysis for banks (Fig 5) assumes the government steps in to ensure delivery of units. The overall costs are similar – either the banks suffer a Rmb 1.1tr mortgage impact or there is Rmb 1.1tr in monetary support from the government or other related parties to help with the delivery of units. Obviously the latter helps to stabilize the impact on society as the burden will be borne by the government.
If Rmb 1.1tr of monetary support is required, we don’t expect the banks to be fully responsible, as this may lead to more developers hopeful of banks coming to their rescue in the future. However, given that the affected developers are already running out of cash, it may be hard for them to take responsibility to deliver the housing units. The CBIRC has also mentioned that they will provide guidance for the problem to be solved in a more “marketization” way. If banks have to digest all the cost by themselves, they are unlikely to be willing to do so as a market player and may even choose to bear this as NPL as it is a similar amount.
Our base case assumption would be banks to take 30% of the monetary support required. Banks should be willing to take this if required by the government, as otherwise they may face Rmb 1.1tr of risky mortgages. As we expect the monetary support to be spread out over a few years, we have taken 3 years in our base case assumption, given that home buyers would be willing to wait if their flat can be delivered.
Expect 4-5% net profit impact in FY22-24F in our base case on unchanged provision coverage ratio.
In our base case scenario, banks will take 30% of the monetary support required of Rmb 1.1tr and spread this over 3 years. This implies more costs of Rmb 111bn per year. Meanwhile, we do not expect NPLs to be significantly affected in this scenario if flat delivery is largely guaranteed. In this case, we estimate 4-5% negative impact on banks’ earnings each year in FY22-24F (Fig 6). The CET 1 ratio would be largely unaffected.
No earnings impact if buffer from provisions is factored in.
Here we assume the coverage ratio remains unchanged at c.200%, while in reality, more provisions might be released as write-backs in this situation and the impact would much lower. We expect c.Rmb 6tr of provisions to be made in FY22F assuming an unchanged NPL ratio as of FY21 and an unchanged 200% coverage ratio. In fact, the provisions should be able to absorb c.Rmb 1tr of additional NPLs if the coverage ratio is lowered to 150%. If so, there is no impact on earnings in banking system even if all the Rmb 1.1tr of monetary support is handled by the banking system in the next 3 years.
A worst-case scenario: Banks to bear all costs.
This is very unlikely in our view, however investors might be concerned with the worst case scenario if banks have to bear all the losses by themselves. Based on our scenario analysis, we expect c.9.5% negative impact on FY22F net profit. The impact is mainly from 1) drop in net interest income; and 2) higher provisions. Here we assume provision coverage ratio to drop from 200% to 150% to release more buffer to digest the NPLs.
Selling to AMCs could be one of the choices.
Even in the worst-case scenario, we did not factor in the possibility of the banks selling their NPLs to AMCs (asset management corporations) if the amounts are too large. This way, banks may sell their loan portfolios at a discount of c.40-50% based on past experience and get a cleaner balance sheet. If this is the case, the impact on earnings is not material, as the provisions may be able to absorb the loss selling the loans at a discount.
What about the impact from Henan village banks?
Besides the mortgage suspension issues, some investors are also concerned whether the Henan village banks’ scandal will affect other banks as well. Again, we don’t expect this to spread or trigger systematic risks. The cause of the event is that some village banks in Henan were doing internet financing and “P2P”-like lending, going against the regulation that rural banks should only do local business. These banks were in fact undertaking national business through internet and the risk is beyond their management’s capability.
We don’t see this event triggering systematic risks. The affected amount was c.Rmb 390bn while total deposits in China’s banking system was Rmb 246tr as of May 22. Currently, we don’t see any signs of the risk spreading to other major banks, as this was more of misbehavior of individual banks, rather than a weak balance sheet in the system.
According to the China Financial Stability Report 2021, PBOC concluded that the 24 large banks in China were all well within the safety margin. There were 271 and 122 rural banks/credit co-operatives and village banks that are recognized as high risk, representing 90% of all high-risk institutions.
Larger SOE banks are very unlikely to be affected and may even benefit from more deposits being transferred from smaller banks to SOE ones. According to our channel check, we do not see high possibility of larger SOE banks taking over or bailing out the affected village banks.
Which banks will be more affected?
Looking at the numbers only, the retail-oriented banks such as PSBC (1658 HK) and CMB (3968 HK) have higher exposure to mortgages among SOE banks and joint stock banks respectively. However, there are factors such as provision capability, exposure to different regions and property developers that we have to take into consideration when assessing the risks of individual banks.
What’s more important than the overall exposure value is the city and property developer exposure. Though it’s hard to obtain mortgage exposure to each developer’s projects, we could use the developer loan as a rough estimate, as the developer’s loan and mortgage for pre-sale of first-hand flats are usually closely linked with each other. Within the first 100 affected project names published, Evergrande accounts for c.30% and 30% is from Henan province. If we use the exposure to Evergrande and Central China region as a rough indicator, CEB (6818 HK) seems to have the highest exposure to risky mortgage loans among our covered names, followed by CITIC (998 HK) and ABC (1288 HK) while the retail banks like PSBC, CMB and PAB have relatively lower exposure.
Higher provision coverage ratio of retail banks provides more cushion.
In fact, retail banks like CMB (3968 HK), PSBC (1658 HK) and PAB (000001 CH) reported higher provision
coverage ratio than peers. As of FY21, PSBC and CMB reported more than 400% provision coverage. The high provision helps to absorb any potential losses before hurting the banks’ net earnings. If we use our base case scenario assumption for the banking system (e.g. 2.9% of the outstanding mortgages would be affected), the provisions could absorb all the losses for these three banks if they turn to a lower coverage ratio of 150% (Fig 10).
Unlikely to affect dividend payout for most of the banks.
Some investors are concerned whether banks are still able to pay out dividends if they suffer huge losses from the mortgage side. In fact, even for those with provisions unable to fully absorb the potential loss, the impact of new NPL formation on Capital 1 ratio is relatively small for the big SOE banks. Their CET-1 ratios will still remain at a high level after the NPL impact and 30% payout ratio is taken into consideration. On the other hand, relatively weaker joint stock banks like CITIC and CEB may have CET-1 ratio below the minimum requirement of 7.5% if they want to keep their payout ratio in the base case scenario of the stress test. CEB may need to seek external capital to meet the minimum requirement.
How will this affect banking sectors’ long-term growth and valuation?
To wrap up, we expect the earnings impact on banks to be limited. A more likely situation will be that the government steps in and shares the costs with the banks. The earnings impact is estimated at 4-5% and could be fully absorbed by provisions. Even in the worst case scenario, the negative earnings impact is manageable at below 10%.
Retail banks should be protected by their strong provisions with little earnings impact, while Big Four and BOCOM would benefit from their strong capital position with no impact on dividend payment. Among all the banks, we see CITIC and CEB have relatively weaker positions but the chances for material losses are low.
The Banking sector was down 2-14% in the past one week (Fig 12). Interestingly, in terms of magnitude of the share price drop, CITIC/CEB < Big Four and BOCOM < Retail banks. The share prices of both PSBC and CMB dropped more than 12% during the last five trading days in the HK market. The possible explanations are 1) these retail names have a high exposure to mortgages if just looking at the numbers; and 2) the retail banks trade at a much higher valuation than peers, while the rest are trading at 0.2-0.4x P/B prior to the drop with limited downside.
Expect share prices to remain under pressure in the near-term.
Based on our scenario analysis on earnings and recent share price performance, we view the correction has priced-in even in the worst-case scenario especially for the retail banks. However, we do expect the sector to remain volatile in the near term. Though we have seen some marginal improvement in June’s presales data, the property recovery process is likely to be delayed as potential home buyers are very cautious now. This will impact mortgage growth and banks’ interest income, which has not been included in our previous analysis. Also, a delayed property market recovery is negative to China’s overall economic
growth, which is also negative on banks’ profitability and asset quality.
From the banks’ perspective, mortgage remains a good product. The weighted average NPL ratio for mortgages for our covered banks show that the NPL ratio was 0.2% in FY21, while that for other loans was 1.78%. At the same time, mortgage’s interest rate is usually higher than 4.5%, and at a premium to corporate loans. The weaker demand for mortgages would thus affect banks’ profitability.
Prefer HK banks to China banks.
At this juncture, we are turning more neutral on the China banking sector, especially in the near-term. Potential positive catalysts are 1) clear and actionable government guidance on how to solve this issue;
and 2) more signals of economic recovery in late 3Q and 4Q. We prefer the HK banking sector to China banking sector as HK banks would benefit from the interest rate upward cycle and have little exposure to China mortgages. Our top pick is CMB is now trading at c.0.9x FY23F P/B, a 5-year low. PSBC is trading at c.0.5x FY23F P/B, close to its 5-year low of 0.4x. We maintain a positive view on China’s economic recovery in the longer term and opine that the banks deserve a higher valuation. In the longer-term, retail loans should still be good assets with high growth potential. The capability of higher retail AUM and operating in a “light asset model” are important to improve the ROE, especially when mortgage growth might be slowing down.
Maintain positive rating on selected retail banks in the long-run.
Still, we maintain our positive stance on the China banking sector, especially on retail banking names like CMB and PSBC in the long-run. The 1H22 bank results should be in line or better than expected, based on a few A-listed banks which have already issued quick overviews. By the end of Jun, the overall NPL ratio in China was 1.77%, or only a slight deterioration compared with 1.76% as reported in 2Q21. The negative impact on asset quality caused by the economic slowdown was not as large as the market
expected. We maintain a positive outlook for China economic long-term recovery trend.