Credit Suisse on Monday joined those bashing China’s useless banks, downgrading the sector to underweight due to “asset quality concerns”.
As ever, definitional and measurement issues ensure that one man’s NPL is another man’s PL. Nevertheless, Credit Suisse’s assumptions of credit growth and credit costs would mean Chinese banks — ICBC aside — are in for a tough time over the next few years.
CS starts with a familiar view of the macro picture: “massive off-balance sheet financing”, visible only in credit-to-GDP ratios (rather than loan-to-GDP ratios), “raises a red flag for future asset quality problems in banks”.
Analysis by our China strategist Vincent Chan (Debt threats, 20 June 2011) shows that during 2009 and 2010, total system credit financing totalled Rmb26.7 tn, representing a 71% jump in total system credit from an estimated Rmb37.8 tn in end-2008 over a twoyear-period. The credit-to-GDP ratio jumped by 46 p.p. from 120% in end-2008 to 166%by Mar-11. …
Using the social financing data released by PBoC for the first time, we find that credit penetration, including the previously opaque off-balance sheet financing, has risen sharply in China over the past two years. In fact, it is already above the critical levels of credit-to-GDP being >10% ahead of the long-term trend which has, as per econometric models of the Basel committee, led to banking sector stresses in many markets (Figure 12-15). This statistic is not visible if we simply use the on-balance sheet lending of the banking system. …
In a bottom-up approach, analysing the 1Q11 cash flow of some 1,000 non-financial companies in China we estimate that interest coverage was 8x, down close to the levels last seen during the Global financial crisis (2008-09).
To wit, contrast:
Red flag noted, CS is no longer expecting a soft landing for China, even if inflation genuinely comes down.
It then tries to estimate the extent of (and trends in) non-performing loans. A tricky task, as Michael Pettis explained in our recent podcast. Here’s how CS did it:
We had introduced our EBITDA/interest coverage – proxy NPL ratio model in September 2008 (China banks: Asset quality – a risk understated?). We had used listed A-share nonfinancial companies (a total size of about 1,500) as a sample and calculated each company’s interest coverage ratio. We then arbitrarily picked companies with an interest coverage ratio below one and defined loans from these companies as nonperforming. Based on it, we replicate the exercise in 1Q11 but adjusted for the following: 1) interest expenses in 1Q11 at about 30% of FY10 on about 18-19% YoY system loan growth and about 12% increase in interest rate stemming from four rate hikes since Oct 2010; 2) seasonal adjustment for EBITDA, as Chinese companies’ cash flow shows a strong seasonality. Our base-case scenario was based on 1Q11 EBITDA as 50% of full year quarterly average (versus 48%, 73%, 45% and 58% in 1Q07, 1Q08, 1Q09 and 1Q10 respectively).
And here’s what they find:
The charts below (Figure 16 and Figure 17) summarise the results of our exercise. We suggest investors to focus on the trend rather than the absolute amount. Comparing our NPL ratio proxy with the actual system NPL ratio doesn’t seem to help in verifying our NPL ratio proxy, as the system NPL ratio has been falling and staying low at about 1% in 1Q11, but the proxy NPL ratio will probably jump to the 2008 level. We reckon that ample liquidity derived from both easy credit and a bulge in B/S funding in 2009-2010 should have 1) “dressed up” corporates cash flow statement and 2) and made ever-greening easier. But going forward, the system NPL ratio may already hit an inflection point with a softening macro outlook coupled with a tight monetary policy in place, in our view.
This take on NPLs leads Credit Suisse to estimate that credit costs are set to soar for Chinese banks — to levels seen during the Asian financial crisis.
This leads to our credit cost assumption increasing to approximately 75 bp in 2011E (from ~60 bp earlier), 100 bp in 2012E (from ~70 bp) and 150 bp in 2013E (from ~75-80 bp). These levels are not outrageous or theoretical – Thai/Indo/Korean banks lost double-digit percentage of loans during the Asian Financial Crisis (similar case of fixed asset investment binge led by mis-pricing of cost of capital). China itself has seen 200 bp levels in the earlier part of the last decade (based on our checks and estimates, since banks did not start getting listed until 2005). Of course, some observers will worry about evergreening by banks and that such NPLs may never show up. If the GDP downturn is short lived then banks may be able to gloss over it but a protracted slowdown, especially with stricter CBRC (not to forget a new administration taking over in 2012), may lead to better transparency.
Here’s hoping. Until then CS is down on banks:
We downgrade China banks sector to UNDERWEIGHT and ABC to UNDERPERFORM,BOC to NEUTRAL (from OUTPERFORM). ABC has greater risks from its county portfolio,while BOC grew its loan book massively (particularly domestic loans by 56%) in 2009.CCB remains OUTPERFORM on fundamentals but we recognise the technical overhangto be created by the 10.2% stake held by Bank of America coming off the lock-up inAugust. ICBC remains our top pick.
And in case you felt the assumptions about credit costs and asset quality were overly harsh, CS has this useful reminder, buried in page 16.
Chinese banks do have a track record of high levels of impairment as well as provisions in the past, especially before being cleaned up by the government in early part of last decade. While the pre-listing data is not available, our discussions with banks and rating agencies indicate that as much as 20-25% of loans were bad at that time.
There’s much more in the note, which is available in the usual place.