What is the true size of China’s debt load, and how fast is it growing? The answer has significant implications for the global economy. Global watchdogs including the International Monetary Fund and the Bank for International Settlements (not to mention this blog) have become increasingly shrill in their warnings that China’s rising debt load poses global risks.
Estimates of Chinese debt based on official figures are frightening enough — an FT analysis put the figure at 237 per cent of GDP at the end of March — but an increasing number of analysts now believe that the true figure may be higher.
The reason is not, as many China data skeptics would assume, that China’s government statisticians are intentionally cooking the books. Instead, financial engineering and rising complexity in the shadow banking system are outstripping the ability of traditional indicators to track debt flows from all sources.
In focus is a once-obscure data series that tracks bank lending to non-bank financial institutions (NBFIs). Standard debt-to-GDP ratios normally exclude borrowing by banks and NBFIs because the money that financial institutions borrow is on-lent to non-financial borrowers. That means including financial borrowing in a total debt estimates normally leads to double counting.
But as ever, the situation in China is murkier. Bank claims on NBFIs — shown in the People’s Bank of China’s depository corporation survey as “Claims on Other Financial Institutions” — have increased massively, from to Rmb11.2tn at the end of 2014 to Rmb25.2tn at the end of August. This chart from UBS’ Wang Tao shows the growth:
That has raised suspicions that banks are using NBFIs like trusts, securities brokerages, and fund management companies as intermediaries in complex transactions where the ultimate counterparty is a traditional corporate borrower. The result is “hidden credit” not captured in the main data series designed to track credit to the non-financial sector, known as total social financing (TSF), sometimes also translated as “aggregate financing to the real economy.”
The PBoC created the TSF series in 2011 as a way to track credit from the rapidly expanding universe of shadow banking. TSF explicitly excludes government bonds, but it is supposed to capture all credit to the corporate and household sectors. TSF includes a large share of non-bank and off-balance-sheet lending, notably corporate bonds, trust loans, entrusted loans, and undiscounted bankers’ acceptances. But the cat-and-mouse game between banks and regulators has progressed since 2011, with banks pioneering new structures designed to evade regulations.
Bank financial engineering in partnership with NBFIs is especially popular among small and mid-size lenders that are now the fastest-growing players in China’s financial system. Loans channeled through NBFIs are a key source of assets underlying wealth management products, which small banks hold both on- and off-balance sheet.
Loans routed through NBFIs are known as “channel business” (in Chinese, tongdao yewu). When held on-balance-sheet, they channel loans are typically listed as “investments” rather than loans. Investment assets carry lower risk weightings than loans to non-financial corporates — sometimes as low as 25 per cent, versus 100 per cent for a traditional loan. Loan-loss provisioning requirements are also lower.
On-balance-sheet WMPs account for rapid asset growth at mid-sized like Industrial Bank and Shanghai Pudong Development Bank, classified as “joint-stock banks” by China’s banking regulator, and “city commercial banks” like Bank of Nanjing and Bank of Jinzhou.
How much hidden credit?
That’s the background. Now on to the numbers. What is the scale of hidden credit disguised as loans to NBFIs?
Chen Long of GaveKal attempts to answer this question by comparing bank credit to NBFIs as shown in the PBoC data survey to separate industry association data that details how trusts, brokerages, and fund management companies invest the money they receive from banks.
Mr Chen reckons that of the Rmb22.1tn in outstanding bank credit to NBFIs at the end of March, Rmb14.6tn flows to the real economy, of which Rmb5.6tn is either already captured in TSF or is invested in interbank assets that don’t represent credit to the non-financial sector. That leaves Rmb9tn, or about 40 per cent of bank credit to NBFIs, as hidden credit not captured either in TSF or in government bond data. The flowchart from GaveKal shows how funds flow from banks through NBFIs to either securities markets or the real economy:
Wang Tao, China economist at UBS, has a higher estimate of hidden credit than Mr Chen, at Rmb16.7tn at the end of 2015. There are multiple differences between Ms Wang and Mr Chen, some fairly technical. But the biggest single factor is that Ms Wang includes Rmb4.7tn in trust assets that are not classified as loans but which she believes are actually “shadow loans in disguise”. These include so-called equity investments that are more like de facto debt, often because they are paired with repurchase agreements.
Despite the technical disagreements, both Mr Chen and Ms Wang agree on the big-picture conclusion. China’s “hidden credit”, while real, amounts to only a minor adjustment in the context of explicit, non-hidden credit of around Rmb163tn at the end of March, according to the FT’s estimate.
The graph below from GaveKal shows how the addition of hidden credit affects the growth rate of overall credit in recent years.
“Private sector credit” refers to TSF. The grey line shows adjustments to TSF by Jonathan Anderson of Emerging Market Advisors. Like Mr Chen and Ms Wang, Mr Anderson also believes that bank credit to NBFIs is being used to conceal credit now reflected in TSF. But Mr Chen believes that Mr Anderson overshoots by adding the entirety of bank credit to NBFIs to TSF, resulting in double counting.
The black line is the metric that Mr Chen favours. Yet once Mr Anderson’s alleged double counting is removed, the difference between the black line and the blue one — which represents a simple estimate based on official data — isn’t dramatic.
Mr Chen estimates that adding hidden credit pushes China’s total debt-to-GDP ratio from 231 per cent to 244 per cent and growth in credit outstanding in the first quarter from 16 per cent from 18 per cent. He writes that “the trajectory (of credit growth when hidden credit is included) … is not very different from the simple sum of private-sector credit (i.e., TSF) and government bonds, and for most purposes that will still be a fine indicator.”
Ms Wang agrees. Her adjustment for hidden credit adds about eight percentage points to China’s total debt-to-ratio and raises credit growth to 17 per cent in 2014 and 2015 from 14.6 per cent and 15.2 per cent before the adjustment. She nicely sums up the bottom line from the analysis of hidden credit:
All these (sources of hidden credit) do not materially change our assessment that 1) China’s overall leverage is high; 2) credit is growing at a worrisomely rapid pace; and 3) given its unique features of high savings rate, domestic nature of the debt, and large government ownership and influence, China will unlikely face a debt/banking crisis or abrupt stop of the credit cycle for at least another few years.
Cat chases mouse
China’s regulators and the PBoC’s statisticians have not publicly admitted the existence of gaps in TSF. A source close to the central bank told the FT that “there could be some but not a lot” of corporate credit missing from TSF because it is disguised as loans to NBFIs. But there are also signs that regulators are taking steps to crack down on regulatory arbitrage. In July, the China Banking Regulatory Commission published draft regulations that seek to curb the use of brokerages and fund management companies as intermediaries.
When channel lending really took off in 2011, trusts were the main non-bank players that cooperated with banks. When subsequent regulations limited the ability of trusts to play this role, the action shifted to brokerages and fund management companies. The latest rules are likely to reverse that trend, essentially specifying that trusts are the only type of NBFI that can cooperate with banks in channel lending, albeit in a more limited form than in the golden era of 2011 to 2013.
In theory, the result should be a diminution of channel lending and hidden credit. But this being China, much will depend on enforcement, and the cat-and-mouse game will no doubt continue.
Related links:
Regulating those Chinese Weapons of Mass Ponzi – FT Alphaville
Document 82 and slapping down China’s shadow loan market – FT Alphaville
