Parts of China’s “shadow banking system”—broadly defined as the non-bank credit and funding markets—have some troubling similarities to US securitization markets circa 2007–08. Specifically, the “guaranteed returns” and off–balance sheet nature of certain products—as well as a “borrow short and lend long” asset-liability mismatch—are worrisome features, particularly considering that the area, which barely existed in 2008, now has estimated assets between 30 trillion and 35 trillion yuan ($4.8 trillion to $5.6 trillion), or roughly half the country’s GDP. Trust loans, which are used to lend to some of the riskiest industries, are the most “shadowy” part of the system. In this brief we focus on the risks posed by bank wealth management products (WMPs), a pass-thru vehicle where retail customers’ savings are used to provide funds for trust loans (among other things), and where the buyers of such products have little understanding of where the funds are going.

Wealth Management Products: A Ponzi Scheme?

In the WMP business, banks sell short-term investments (WMPs) to their retail clients using a trust or securities company as a conduit to invest the money. Once retail funds are processed through the WMP channel to a trust or securities company, they are often commingled with funds from other WMPs and then invested in a spectrum of risk assets including time deposits, bonds, loans, equities, and even other WMPs. This pooling effectively diversifies the exposure that retail investors face—when one specific investment performs poorly, excess returns from a separate investment can cover the shortfall, allowing retail customers to receive their expected return.

Banks generally market WMPs as higher-yielding, low-risk alternatives to deposits. The average investment term is between one and six months and savers are promised returns several percentage points higher than deposit rates. With inflation at times running higher than deposit rates, these products have attracted retail investors, who often presume the bank will guarantee the return. In reality, very few WMPs have legally enforceable provisions to backstop the investments should they underperform. WMPs are off–balance sheet products and do not show up on banks’ financials statements. Since 2009, this business has grown at a roughly 65% annual rate and is estimated to be as large as 10 trillion to 11 trillion yuan.

The structure of WMPs creates very little transparency for investors. A Reuters analysis of 50 different WMPs could ascertain the final investments for just two. Opacity does not necessarily imply poor asset quality, but marginal businesses with fewer borrowing options seem like an ideal trust loan target. For example, banks are highly restricted in lending to the real estate industry and local government infrastructure projects due to their perceived riskiness. Trusts have had no qualms lending here, and these two industries account for 35% of trust assets.

In an attempt to rein in some of the risks, Chinese regulators mandated in April 2013 that trust WMPs could not invest more than 35% in investments not listed on an exchange, meaning nearly two-thirds of assets would be in “safer” investments. Banks appear to have found an easy workaround—instead of funding WMPs through trust companies, banks now work directly with securities companies. Securities companies cannot make direct loans with their assets, but they can work with trust companies to make trust loans. From mid-2012 to year-end 2013, securities companies’ assets under management exploded from 480 billion yuan to 5.2 trillion yuan. Various sources suggest that the vast majority of these inflows have been through the WMP channel. And as securities companies are regulated by a different authority than trusts, they are not subject to the 35% limit on assets not listed on an exchange. Tellingly, within two months of the crackdown on trust company investments, higher-yielding (i.e., riskier) WMPs proliferated, likely funded through the less regulated securities company channel.

The large duration mismatch between a WMP’s assets and liabilities is another risk. WMPs are short-term vehicles with maturities typically around three to six months, while the assets they fund (e.g., loans for property developers and infrastructure projects) tend to have much longer maturities. To prevent this asset/liability mismatch from causing problems, trusts and securities companies constantly roll their funding, issuing new WMPs to replace existing ones as they mature. New issuance of WMPs exceeds 200 per day, and an estimated 5.3 trillion yuan of trust products will come due in 2014 alone. Pooling WMP funds helps to mitigate the asset/liability mismatch, as new WMP funds can be used to manage WMP redemptions on time. However, looked at another way, when new money is used to pay off old investors, “to some extent, this is fundamentally a Ponzi scheme,” as former Bank of China Chairman Xiao Gang put it in a 2012 China Daily op-ed.

Could WMPs Cause a Credit Crunch?

The US shadow banking system (namely securitization) was a precipitating factor for the 2008 crisis, and we see some parallels in China. First, retail investors in WMPs are widely under the impression that their funds are guaranteed yet have little transparency into where their funds go. This is reminiscent of the buyers of CDOs rated AAA, who ultimately had little idea of the true nature of their investments. Second, the asset/liability mismatch is precisely how the US structured investment vehicle crisis erupted—short-term funds could not be rolled over. Subprime securitization became the channel through which presumably safe institutional money was funneled into risky loans. Subprime lending grew and credit standards diminished to the point that when rates rose and home prices flattened, the market collapsed. Could China be in for a similar fate?

Risks appear to be rising. Retail money flows in at ever-increasing rates, trusts have large exposures to risky industries, and rapidly rising yields indicate that credit is worsening. Staying one step ahead of regulators, securities companies have grown their assets tenfold in 18 months. Add in the asset/liability mismatch, and there’s a powder keg waiting for a match—of which several could be lit. If losses from trust investments flowed through to bank customers, skittish retail investors may pull out of the WMP business en masse, creating a credit crunch. Another spark could be the trust companies. With just 2.3% capital to assets, they are leveraged over 40 to 1, double their 2010 level. Could defaults in their portfolios quickly eat through this capital layer, causing solvency issues? A third potential match is the coming refinancing wave in conjunction with the continued regulatory crackdown. If regulators succeed in lowering WMP issuance rates, will trusts be unable to roll their funding for long-dated assets? Notably, China’s two largest state-owned banks recently stopped distributing trust products.

Given the huge political consequences if meaningful losses are forced onto retail savers, trusts and sponsoring banks will likely be left to work out bad loans and absorb losses. The size of the potential hit is difficult to gauge, but trust assets are roughly double the Tier 1 capital of major Chinese banks. Unexpected significant stress in the banking system would be a substantial headwind to China’s growth, as credit availability would decline across the board, and likely be unavailable at the margin. While retail savers would be made whole, the WMP channel would likely close down, much as subprime securitization did during the credit crisis. At about 15% of GDP, the loss of this credit channel could be substantial. Industries seen as more reliant on shadow lending, such as infrastructure and property development, may be hardest hit. This, along with known overcapacity in industries such as coal and steel production, would put continued downward pressure on commodity prices. China’s banking system is relatively isolated from the outside world, so even in a severe credit crunch it would be unlikely to create the systemic risk of a global financial crisis. However, if China significantly undershoots estimates for economic growth this year and next, global growth could come under pressure. Emerging markets would be especially at risk, as would commodity exporters, but so too would global equities in general.

What to Watch

The China Banking Regulatory Commission is now requiring trusts to develop capital-raising contingency plans for a liquidity squeeze. Trust companies must submit plans by June 30 outlining their strategy to unwind non-tradable assets. How well developed and realistic these plans are could provide investors with a better gauge of just how much systemic risk exists in this part of China’s shadow banking system.

The Bottom Line

China has undergone a record-setting credit boom since 2008, with retail deposits filtering through lightly regulated lending channels into industries where banks have been restricted. The lack of transparency, misunderstanding of credit risks, and asset/liability mismatches all combine to heighten the risks. A credit crisis in China is not a given, but global investors’ low expectations of one seem disconnected from realities on the ground. Regulators’ recent crackdown on WMPs—an attempt to remove risk from the system—could add to it if the lack of new WMPs for rolling funds creates cascading defaults in existing WMPs.

Contributors

Tim Hope, Manager, Analytics
Emily Ginsberg, Investment Associate

Exhibit Notes

Estimated Size of China’s Shadow Banking System
Source: Standard Chartered.
Notes: Data on China’s various credit channels vary and these estimates should be viewed as such. Estimates for trust loans includes loans made by trust companies as well those made by securities companies.

Estimated Number of WMPs Issued Weekly by Expected Returns
Sources: BofA Merrill Lynch and Wind Information Co., Ltd.

Estimated Average Number of WMP Products Issued per Day in Each Quarter
Sources: BofA Merrill Lynch, Puyi Wealth Management, and Wind Information Co., Ltd.
Notes: Data from June 2013 to December 2013 are interpolated from weekly data from BofA Merrill Lynch and Wind Information Co., Ltd. The data are aggregated by quarter and divided by the number of business days each quarter (60).