It has become a note of the excessively optimistic China bull to argue China is deleveraging. In part I of this brief series, I addressed the specific and narrow data point of non-financial corporates the deleveraging crowd is relying on to argue for deleveraging. I then also focused on the rapid rise in household debt that now amounts to 102% of household income and is rising rapidly.
To arrive at the deleveraging argument, which is not happening even using their own data, they omit multiple sectors of the credit markets. One of the biggest that is overlooked, and very commonly overlooked by many people, is the growth in credit to the financial sector. This is not a trivial matter.
A few years ago when McKinsey Global Institute released a study on Chinese debt levels it was criticized by some as over estimating the level of Chinese debt because it included debt owed by the financial sector. The logic is this: if a bank borrows money to lend to a real estate developer or a coal mine, there is no fundamental difference than them accepting deposits to lend to a real estate developer or coal mine. In short, many claimed counting debt owed by the financial sector was double counting debt.
In fairness, there is some merit to this argument. Think of two simple examples where this is a good argument. First, assume there is a bank and any company in the real economy. If the bank just borrows money to lend to the other company, maybe in a combination of bonds or as deposits, then we would be double counting if we counted both the financial institution debt and the other companies debt.
Second, think of an economy with two banks and a company in the real economy. Assume the real economy company asks for a $100 loan. The loan doesn’t get made because Bank A has $70 in deposits and Bank B only $51. Bank A can’t make the loan by itself and Bank B doesn’t feel comfortable using almost all its capital. However, Bank A offers to loan Bank B some money say $9, at a rate slightly above its deposit rate, and the $100 loan gets made to the real economy company. Again, if we count the financial debt, we would be double counting the actual debt outstanding. The key is that the financial debt nets out in the real economy between banks.
Financial debt should not be netted out however when it does not flow into the real economy and merely shifts between financial institutions increasing leverage. Assume there is one real economy asset owned by Party A. Party B thinks it will appreciate in value in borrows 90% of the money needed to purchase the asset and buys it for a slightly higher price. Party C thinks it will appreciate in value also and borrows 90% of the money needed to buy it from B and pays a slightly higher price. Repeat as needed. Here, there is only one underlying asset but debt has increased rapidly. As long as the price continues to appreciate, everyone is happy and makes money.
Anecdotally, I can personally attest to more than a few cases, where in one instance, someone owns an apartment free and clear or with virtually no balance relative to value. They take out a loan pledging the apartment as collateral using proceeds to purchase another apartment free and clear. They then go to another lender pledging the second apartment as collateral using those proceeds to purchase a third apartment. Officially, each lender has made a loan into the real economy with high grade collateral. In reality, financial debt has tripled and only one real asset exists.
The problem with financial debt, as noted in the example above, it is not always clear even for specific lenders much less at a more macro level to disentangle financial and non-financial debt.
If we look at the monthly balance sheet of “Other Depository Corporations”, referring to banks besides the PBOC, and their claims on financial institutions the data is very revealing. Since February 2016, depository corporation claims on other depository corporations (read mainline traditional commercial banks) are up a paltry 3.9%. However, claims on other financial institutions are up 29.1% and only trail total claims on DCs by 3.1 trillion RMB out of 31 trillion in claims on other DCs. At current rates of growth, even allowing for some slowing, claims on other FI’s will over take claims on DCs sometime this year.
For accounting and capital adequacy reasons, which have been well explained elsewhere, much of the financial-to-financial flows that have taken place have not taken place officially as a loan. Whether it is in a “negotiable certificate of deposit” or “investment receivable”, much of these flows which are effectively credit instruments are not labelled as such.
In fact, if we look at the monthly data on sources and uses of funds of financial institutions, we see that “loans to non-banking financial institutions” was up a relatively modest 8.9% totaling a pretty modest 800 billion RMB. However, if we include the category “portfolio investment” which is comprised of “portfolio investments”, and “shares and other investments”, we receive a very different picture. Portfolio investment is up 33% from February 2016 to February 2017. In other words, according to the official uses of funds data, loans to non-bank financial institutions barely changed in relative and absolute terms. The “other” categories though exploded.
To put this in perspective, the total increase in the total use of credit funds in financial institutions in the household and non-financial sector from February 2017 to February 2016 was 13.2 trillion RMB growing by a moderately robust 12.9%. These three categories of use of credit funds by financial institutions however grew by a total of 24.1 trillion RMB. In other words, use of credit funds at financial institutions to “financial sectors” grew 82% faster than lending to firms, government, and households.
Even wealth management product statistics indicate that only about 60% of wealth management capital goes into the real economy, and even this number should be treated cautiously. With the allocation of WMP capital into commodities (read coal and steel among others) recently tripling from a year earlier which was the driver of price changes not changes in supply or demand, it is clear these financial to financial flows are having a major distortionary impact on the economy and should not be treated as simple double counting as in the simple early examples.
All this leads to a couple of conclusions and scenarios. First, growth in financial related debt is rapidly outpacing growth in debt in the real sector (non-financials, governments, households) in both relative and absolute terms. This is worrisome for what it possibly indicates and how investors view the state of the overall economy. In short, there simply are very few good projects, even by lax Chinese socialist market conditions, and banks would rather hold financial assets.
Second, one implication is that there is likely a significant upside deviation in the true value of asset prices. This flood of financial capital into asset markets pushing prices higher appears to reflect the reality of the Chinese economy. From tech start ups to real estate to coal, valuations and speed of price changes bear little resemblance to underlying fundamentals. As a point of comparison, the 100 City Index average in December 2016 was 13,035 RMB/sqm while the urban per capita disposable income was 33,616. Through some additional calculations, this yields a home price to household income ratio of 14. The evidence seems to bear out the idea that asset prices are inflated.
Third, it also implies that banks simply do not have the necessary liquidity needed potentially indicating greater credit risk than is being acknowledged. PBOC balance sheet claims on depository corporations are up 68% from February 2016. The explosion of financial debt and rapid increase in central bank holdings is telling us banks are liquidity constrained. Given that financial debt has grown so much more rapidly than non-financial and household, this would imply that financial institutions and firms are being propped up to avoid significant problems.
Fourth, the reality is probably a mix of these two things as asset prices and debt are so intertwined in China that the only way things work is by keeping asset prices going up. The problem here is that this continues to increase the multiple layers of hidden leverage. Like the example where someone owns an apartment then borrows against it, buys another, borrows against it, then buys another, as needed, this makes any decline in real estate values potentially suicidal.
Want an example of this collision of increasing financial debt propping up asset prices? Major financial institutions are setting up subsidiaries or related companies that issue debt to buy bad loans at face or near face value. They are even issuing bonds to buy bad loans. Rather than acknowledging some significant loss taking a hair cut and lowering the asset value, they are reissuing debt to buy bad assets at near face value.
Given the weight of evidence, not only is China not deleveraging, its financial debt is rapidly outpacing its growth in real economy debt growth providing worrying signs about the state of Chinese finances. It also tells us it is a major mistake to simply deduct all financial debt.
Side note: How worried is China about the rise in financial and interbank debt? Just a few days ago, Caixin had three articles about this on the cover of its landing page though with different dates. Article 1, Article 2, and Article 3.