While the world has started paying attention to China due to its stock market having the stability of your average reality star, I have always cautioned and continue to believe that whatever gyrations the Shanghai and Shenzhen indexes experience is a mere sideshow. One of my primary areas of concern is the credit market.
I think there is good reason to be worried about the state of the Chinese credit market that move well beyond the standard explanations of excessive credit growth and a slowing economy, which are perfectly valid by themselves. RBS Albert Gallo via FT Alphaville writes that “we estimate peak NPL’s of 5.5% in China” using I believe a reasonable and sound methodology for estimating how high NPL’s will increase from their current state. I think the RBS method based upon credit in excess of credit to GDP trend is a reasonable estimate of how much on average we could expect NPL’s to increase from their current state.
Here is where I differ, using the current official data on NPL’s in China enormously understates the number of bad loans in China. This is not some conspiracy theory or complicated statistical analysis, the banks are literally telling you their definition of an NPL does not resemble anything that would be acknowledged outside of China. As one paper from economists at Wharton(PDF) wrote in a recent paper:
“…the classification of NPLs has been problematic in China. The Basle Committee for Bank Supervision classifies a loan as “doubtful” or bad when any interest payment is overdue by 180 days or more (in the U.S. it is 90 days); whereas in China, this step has not typically been taken until the principal payment is delayed beyond the loan maturity date or an extended due date, and in many cases, until the borrower has declared bankruptcy and/or gone through liquidation.”
Nor is this a case of conspiracy theories by Chinese and foreign economists. The Chinese banks actually tell you the same thing. In recent bank IPO prospectuses when banks provide a wealth of information such as loan classification systems, they would say essentially the same thing. Huishang Bank for instance, wrote that a major regulatory risk is that “our loan classification and provisioning policies may be different in certain respects from those applicable to banks in certain other countries or regions.”
Amusingly, the next risk identified by Huishang is that “we cannot assure you of the accuracy of facts, forecasts, and statistics derived from official government publications contained in this prospectus with respect to China, its economy, or its banking industry.” In other words, Chinese banks don’t believe official statistics.
Banks that went public in 2014 took risk recognition two steps further. One bank Shengjang wrote that “our historical non-performing loan ratios may not fully reflect the actual changes of our asset quality due to government-sponsored disposals and write-off of non-performing loans in the past that were not in the ordinary course of business….we cannot guarantee that the government will continue to help us dispose of and write off our non-performing assets.” In other words, their NPL ratios were protected due to government bailouts and they can’t guarantee this will continue.
However, it doesn’t end there. The Chinese banks actually say they might not have data on even who they lend money to or any documentation to support their claims (no, I’m not making this up). Shengjing actually writes in their IPO prospectus ““loan is unrecoverable despite a favorable court judgment due to our failure to apply to the court for enforcement before the applicable deadline” or “no loan contract (agreement) has been signed with the borrower, or the original loan contract (agreement) has been lost, and the borrower refuses to confirm the loan…”. In other words, even if they can get a favorable court judgement, they might not have the paper work or information on the loan. Questions about China data from firm to national level are not conspiracy theories but stated bluntly by firms and political leaders themselves.
All this leads into the two specific reasons this matters and why the RBS estimate of 5.5% peak China NPL is an underestimate in my opinion. First, there is a very high degree of probability that we are already at 5.5% with numbers continuing to rise rapidly. Special mention loans plus official NPLs in China are already at 5.13% and has risen 43% in the past year. Given what we know about Chinese loan classification, which Moody’s noted recently going even further noting that loan classification quality was actually deteriorating, it does not stretch credibility to believe that 5.5% is already in sight. The rapid rise of “special mention” loans, the classification before a loan becomes NPL, indicates that banks are trying to avoid reclassifying them which coupled with the rapid rise indicates significant unreported stress.
Second, a significant amount of lending is being conducted in increasingly shorter duration. For instance, Harbin Bank saw nearly two-thirds of its loans in short term holdings under one year and continuing to rise. The shift and continuing rise on short term lending would seem to indicate stress in repayment. All evidence points to mid-size and smaller banks along with shadow lenders having large portfolios of short duration assets under one year and the 4 majors having about 1/3 short term lending with longer term lending flowing heavily to SOE’s and similar type firms. This means that every year large amounts of debt needs to be rolled over or paid off. With these numbers continuing to rise as portion of total debt this seems to indicate unreported stress in the credit market. Couple this with what we known about debt classification and it paints a worrying picture.
While I think the RBS analysis is solid, I don’t believe it is starting from the right base. If we believe that a not insignificant portion of special mention loans will migrate into NPLs, which is not an unreasonable assumption given a slowing economy then adding in a belief about loan classification with Chinese characteristics and we can already see 5.5% on the horizon.
Note: As usual,here is a link to the NPL dataI reference downloaded from WIND.Here is a link to a paperI wrote on Chinese banks, NPLs, and repo lending. Anyone that wants the referenced IPO prospectuses just email.