How China's Banks Hide Trillions In Credit Risk: Full Frontal

Tyler Durden's picture

On Monday, we noted that NPLs in China saw their biggest quarterly increase on record in Q1, jumping 141 billion yuan sequentially to 983 billion. We also discussed why the proliferation of property loans is spreading real estate risk to the larger economy before outlining several reasons why the “official” data on bad loans may severely understate credit risk in much the same way that “official” data out of Beijing on economic output probably grossly overstates GDP growth. As a reminder, here is what we said:

One might be tempted, upon reading this, to point to "official" data on bad loans at Chinese banks on the way to concluding that "modest loss-absorption capacity" or not, sub-2% NPL ratios certainly do not seem to portend an imminent catastrophe. However, given the official figure is just 1.39%, and given what we know about the state of China's economy, one could be forgiven for wondering if NPLs at China's biggest lenders are grossly understated. To let Fitch tell it, determining the true extent of China's NPL problem is complicated by a number of factors and the 'real' data might be just as hard to get at as an accurate reading on Chinese GDP.

In order to truly understand how exposed China’s lenders are to potentially toxic “assets”, it’s worth taking a closer look at all of the different non-standard channels through which the banks have taken on credit risk. 

At the outset it should be noted that it is virtually impossible to determine how many of the traditional loans banks carry on their books would actually be impaired were it not for what Fitch calls “centrally managed NPLs.” Put simply, the government has a tendency to ‘influence’ banks’ decisions when it comes to rolling over problem loans. Here’s Fitch:

It is not uncommon for the authorities (central or provincial government, PBOC, or China Banking Regulatory Commission; CBRC) to encourage banks to roll over loans to support the broader economy. Such influence from the regulators is common in China, but less so in more developed markets as it runs counter to the principle of strict prudential oversight. (In addition, in some cases in China the regulator has lowered risk weights to encourage banks to lend to potentially riskier borrowers.)...

Where a bank agrees to provide additional funding to prevent default, i.e. by rolling over the loans and requiring more collateral as credit enhancement, this exposure and/or other loan exposures relating to the same borrower may not be recognised as impaired. The level of disclosure of widely reported corporate defaults (or near defaults) is low, and there is no information available on the amount of debt renegotiations that are unreported by the media. 

In other words, there’s no way to know how pervasive Beijing’s practice of forcing banks to roll-over problem loans truly is, meaning that even if we ignore the fact that quite a bit of credit risk is obscured by the practice of shifting it around, moving it off balance sheet, and reclassifying it, (i.e. if we just look at traditional loans) it’s still difficult to know what percentage of loans are actually impaired because it’s entirely possible that a non-trivial percentage of sour debt is forcibly restructured and thus never makes it into the official NPL figures. 

Fitch goes on to note what we mentioned earlier this week, namely that the percentage of  loans which are not yet classified as non-performing but which are nonetheless doubtful is much higher than the headline NPL figure and in fact, the ratings agency seems to suggest that some Chinese banks (notably the largest lenders) may be under-reporting their special mention loans:

Banks in China have remarkably similar NPL ratios and performance trends, which suggests little or no difference in risk appetite despite large differences in ownership and systemic importance, and varying levels of investment in their risk-management systems. However, state banks’ capital adequacy ratios have clearly benefited from having greater access to capital markets due to their much larger franchises. 


There are more notable differences in special-mention and overdue loan trends, implying that loan classification standards, which in principal are on a par with international practice, may not be applied consistently and uniformly across all banks in China. For example, loans overdue for more than three months are not always classified as NPLs at some banks. The only common trend is that all banks’ provision coverage ratios fell during 2014 as the increase in loan provisioning was insufficient to offset the NPL increases, even as NPLs have been partly offset by write-offs and/or disposals.


Interestingly, it appears that when overdue loans are included, the banking sector’s loan loss reserves aren’t sufficient...

The sector’s provision coverage declined to 232% at end-2014 from 283% a year before, even though the reported NPL ratio only rose marginally to 1.25% from 1.0% over the same period (Fitch-rated banks’ reported NPLs rose another 14% quarter on quarter in 1Q15). This ratio would fall to 66.5% at end-2014 if special-mention loans were included. This means China’s loan-loss reserve, equivalent to 2.9% of loans at end-2014, is insufficient to cover both NPLs (1.25% of loans) and “special-mention” loans (3.1% of loans). 

...but ultimately it’s irrelevant because between bad assets that are ultimately transferred to AMCs, loans that are channeled through non-bank financial institutions and carried as “investments classified as receivables”, and off-balance sheet financing, nearly 40% of credit risk is carried outside of traditional loans, rendering official NPL data essentially meaningless in terms of assessing the severity of the problem:

China’s four major AMCs were set up in 1999 to absorb CNY1.4trn in bad assets at par value from China Development Bank and the big four banks (Industrial and Commercial Bank of China, China Construction Bank, Bank of China and Agricultural Bank of China) before their restructuring. NPL disposals to AMCs have increased in recent years as more banks have come under pressure to manage their reported NPL levels…


Banks work with non-bank financial institutions (ie trust companies, securities firms, other financial subsidiaries and/or affiliates to the banks) to channel credit to sectors that have restricted access to traditional banks loans. This is usually called “channel” business in China. The banks may put together these transactions themselves, in which case the borrower could be an existing bank customer, or the non-bank financial institution may help the bank identify profitable lending opportunities. Some of this credit is sold to investors as WMPs, but banks are increasingly holding it on their own books as “financial assets held under repurchase agreements” or “investments classified as receivables”...


The amount of informal loans channelled through non-bank financial institutions and accounted as “investments classified as receivables” has increased from close to zero in 2010 to CNY4.4trn at end-2014 for Fitch-rated commercial banks, equivalent to 8% of total loans. 


Off-balance-sheet financing (I.e. trust loans, entrusted loans, acceptances and bills) accounted for 18% of official TSF stock at end-2014, up from less than 2% just over a decade ago. Of the off-balance-sheet exposure reported at individual banks, this is equivalent to 15% of total assets for state commercial banks and 25% for mid-tier commercial banks, on a weighted average basis. These ratios would be even higher if we included entrusted loans (see Figure 2), although this information is not disclosed at all banks.

Fitch estimates that around 38% of credit is outside bank loans.


*  *  *

There are several takeaways here. First — and most obvious — is the fact that accurately assessing credit risk in Chna is extraordinarily difficult. What we do know, is that between forced roll-overs, the practice of carrying channel loans as "investments" and "receivables", inconsistent application of loan classification norms, and the dramatic increase in off balance sheet financing, the 'real' ratio of non-performing loans to total loans is likey far higher than the headline number, meaning that as economic growth grinds consistently lower, the country's lenders could find themselves in deep trouble especially considering the fact that loan loss reserves aren't even sufficient to cover NPLs + special mention loans, let alone defaults on a portion of the 38% of credit risk carried off the books.

The irony though is that while China clearly has a debt problem (282% of GDP), it's also embarking on a concerted effort to slash policy rates in an effort to drive down real rates and stimulate the flagging economy, meaning the country is caught between the fallout from a shadow banking boom and the need to keep conditions loose because said boom has now gone bust, dragging credit growth down with it.

In other words, the country is trying to deleverage and re-leverage at the same time.

A picture perfect example of this is the PBoC's effort to facilitate a multi-trillion yuan refi program for China's heavily-indebted local governments. The idea is to swap existing high yield loans (accumulated via shadow banking conduits as localities sought to skirt borrowing limits) for traditional muni bonds that will carry far lower interest rates. So while the program is designed to help local governments deleverage by cutting hundreds of billions from debt servicing costs, the CNY1 trillion in new LGB issuance (the pilot program is capped at 1 trillion yuan) represents a 150% increase in supply over 2014. Those bonds will be pledged as collateral to the PBoC for cheap cash which, if the central bank has its way, will be lent out to the real economy. So again, deleveraging and re-leveraging at the same time.

This is just one of many 'rock-hard place' dynamics confronting the country as it marks a difficult transition from a centrally planned economy based on credit and investment to a consumption-driven model characterized by the liberalization of interest and exchange rates.