Wealth management products have helped bring shadow banking mainstream.

On a certain day in October, 2015, a group of disgruntled investors gathered in Beijing to lodge a complaint. They had bought so-called wealth management products (WMPs)—investment instruments somewhat similar to mutual fund shares—from a company called Global Wealth Investment, which had invested in loans backed by Hebei Financing Investment Guarantee, a state-owned guarantor managing nearly $8 billion in assets which collapsed earlier that year, owing about $98 million to Global Wealth, according to the financial magazine, Caixin.

“We gave great weight to Hebei Financing’s guarantees,” one inve stor told the Financial Times. “We knew it had the backing of the state.”

On the same day, about 200 investors protested outside the Beijing offices of Ping An Insurance, which had sold around RMB 100 million of products from Gold Match Silver Fund Management, whose top executive had disappeared. Ping An, the investors claimed, had swindled them.

So far, such defaults (and the angry investors that accompany them) have been uncommon in China’s complex and under-regulated wealth management product space. But the now-gargantuan industry, which according to Bloomberg comprised $3.9 trillion in holdings as of June 30, 2016, up 11.8% in six months, may pose a large risk to China’s financial system.

This August, the International Monetary Fund issued warnings on China’s credit boom risk, specifically calling out wealth management products as a key part of the problem. China’s overall debt level has risen from about 150% in 2008 to 240% of GDP today, and according to IMF’s statement, the state banks are repackaging at-risk loans into “investments,” which cosmetically improves their balance sheets, but does not fix the core problem—high levels of non-performing loans.

“Economic losses are real whether you recognize them or not,” says Patrick Chovanec, Managing Director and Chief Strategist at Silvercrest Asset Management Group in New York. “Somebody is going to bear it.”

There are many risky aspects of the wealth management products industry in China: use of the money to fuel inefficient or even troubled state-owned enterprises, or to buy property and stocks—both bubble-prone markets in China—cross ownership of WMPs by institutions, and the rolling nature of the system whereby new products help cover the liabilities of previously-issued products. The web of risk, and the scale of it, has analysts like Charlene Chu, of Autonomous, a global equity and credit research firm, worried about the possibility of a chain reaction similar to the 2008 financial crisis when the US mortgage market buckled under similar strains.

“We call off-balance-sheet WMPs a hidden second balance sheet because that’s really what it is—it’s a hidden pool of liabilities and assets,” Chu to Bloomberg earlier this year. “In this way, it’s similar to the Special Investment Vehicles and conduits that the Western banks had in 2008.”

One of the most troubling aspects is the involvement of everyday people across China, who sometimes risk their entire savings in these products. Many of the products have all the trappings of guarantees without a guarantee, and the losers from Global Wealth and Gold Match did not just lose money, they also lost trust. A broader crisis of faith in the system could turn off the money taps from ordinary investors, leading to a liquidity crunch, and likely also widespread anger.

“If they are selling investment products and they are understood to be investment products then that is not necessarily a problem,” says Chovanec. “Except that’s not how a lot of people [understand] it.”

So far, however, most people are still making money, and lots of it. According to China Central Depository and Clearing Co., the first central securities depository approved by the State Council in China, banks and other lenders made RMB 117 billion ($17.3 billion) from WMPs in 2015. But it is unclear how long the good times will last.

“Wealth management product” is a term used to describe off-the-books financing instruments used by banks in China. Buying one is somewhat similar to buying shares in a mutual fund, except the return is fixed, and it usually has a set maturity which is typically short. It usually unclear what the underlying assets are.

Each product is created individually and so terms, interest rates, maturity and other details vary considerably. But generally speaking, on average they cost in the neighborhood of a few thousand dollars, mature in about six months and typically carry an interest rate return of 3-5% (although sometimes the return can reach 10% or greater), which is considerably better than bank deposit rates. For comparison, the savings account interest rate at ICBC (the world’s largest bank by assets) was set at 0.3% in September, and the one-year time deposit rate was just 1.75%.

WMPs are sold directly by banks, and also through third parties. Confusingly, they are also often sold by third parties at bank locations, a distinction that is not always clear to buyers. The proceeds are then invested or re-invested, and everything is game, from the stock market to construction projects.

But the basic concept is simple: “WMPs function as both loan substitutes and deposit substitutes,” says Jack Yuan, associate director, financial institutions at Fitch Ratings in Shanghai. In other words, WMPs are an instrument of shadow banking.

Wealth management products are issued by just about all banks in China, from giant institutions, like ICBC, to mid-tier provincial banks. The largest group of buyers is individual mass-market investors who, according to Bloomberg, in September held roughly half of them.

These products are not inherently exciting, were it not for their enormous scale and the speed of their overall growth in recent years. According to the China Central Depository and Clearing Co., in 2011 they amounted to RMB 4 trillion, and have since ballooned more than sixfold.

“Growth started around 2010, and it was for sure impressive growth every year,” says Denis Suslov, a financial industry analyst at Kapronasia in Shanghai. “2010, ’11, ’12 saw at least 20% growth each year.”

There are factors driving growth on both sides, supply and demand. On the demand side, there are relatively few investment options for ordinary people in China outside of property and the stock market. In 2010 and 2011, when wealth management products really began to take off, the inflation rate was more than 4%. Match that with China’s low deposit rates meant saving money was the same as losing it, and people were happy to have an alternative.

The supply side is a bit more complicated. Chinese banks, like banks in any country, need to turn a profit, but they also need to meet the demands of the state in a system that is not purely commercial. So far, WMPs have been a reliable source of good revenue for them, but also a tool to help them meet their policy obligations. As the government pushes banks to deal with non-performing loans (NPLs), many to state-owned companies, banks can use WMPs to move them off the books. The IMF also warned against this in August.

“Issuing more products off balance sheet is a way for them to make their financial metrics and their balance sheets look a little bit nicer,” says Yuan. “They don’t report NPLs that are in the WMP pools.” But to really understand wealth management products and their place in the system, it is necessary to take a look at the development of banking in China over the past decade.

So-called shadow banking has long existed in China, and it is not inherently a negative. The formal banking system, being a state-owned operation, preferentially lends money to big state-owned enterprises, the consequence being that there are many under-served customers starved for credit, as well as the aforementioned demand for investment vehicles. Shadow banking can fill in the lending gaps for both parties, and to good effect.

But a sea change occurred with the 2008 global financial crisis and the massive Chinese government stimulus that followed. As Chovanec explains, the stimulus loans, funneled largely through the state banks, put huge stress on their books, which in turn caused concern in the government.

“Starting around late 2010, there was an effort by the PBOC to try to rein in that lending, and it involved raising the reserve ratio [among other things],” says Chovanec. However, if economic growth was to continue on target, projects needed to be funded. “What happened was banks figured out ways of making loans besides making loans.”

Those other ways of making loans were wealth management products: effective, profitable and off the books. With these products, banks could improve their balance sheets, and lend with a lot less regulation, including lower capital ratios.

By last year, they had risen to RMB 23.5 trillion, about 35% of China’s GDP—meaning shadow lending is no longer a dark corner of the financial system.

“This is not something that is separate from the banking system,” says Chovanec. “It is an integral part of the banking system.”

Another key event that helped shape the current system occurred in 2012. Originally, wealth management products were used as money market funds, the money market being the liquidity used for short-term borrowing between banks and other financial institutions. However, when money market yields fell in 2012, issuing institutions could not achieve the returns they needed to repay the higher interest on wealth management products.

“The real transformation has been a move to add duration and credit risk to the assets backing WMPs,” says Logan Wright, a director at Rhodium Group who leads the firm’s China Markets Research. Those assets could be something like a housing development, or a coal mine, projects that may not create a return for several years.

According to a statement in September by the China Banking Wealth Management Registration System, about 40% of the WMP funds in the first half of 2016 went into bonds, 17.7% into cash and bank deposits, and 16.5% into “non-standard credit assets” (a hodgepodge of loans)—assets that are longer term. But, good assets in China have begun to dry up, kicking off what Wright calls a “search for yield” that can keep the system moving.

“This is how you have seen the explosive growth of margin financing during the equity markets [boom], the commodities futures volume explosion in April, as well as some of the rally in the corporate bond market right now,” says Wright.

But this is where things get start to get murky—unlike a mutual fund, which has a prospectus, there’s no way to easily figure out what precisely a particular wealth management product is backed by. One Shanghai professional who has occasionally purchased WMPs said she had no idea what her money was funding. Rather, she understood only the basic terms of maturity and interest rate, as well the name of the institution that sold her the product.

“There are just some general statements that there are investment risks and that these wealth management products are not deposits,” says Suslov. “The risk notice is not really adequate.”

Customers mostly buy products from reputable institutions that may only implicitly stand by them—it is not really the bank’s liability, because it is classed as an investment product.

This gets to the heart of the odd economic disconnect inherent in WMPs. In a traditional investment, the investor would assess the risk and price accordingly, knowing there is a chance of losing money. But with WMPs, customers place the bet basically on the institution that is marketing the product, which may be selling an off-the-books amalgam, or something created by a third party.

While few WMPs have so far gone bust, and the state banks pushing them have demonstrated their willingness to guarantee the investors’ principal. According to Reuters, CITIC did just that in 2012 when a wealth management firm marketed by the bank missed a $1.12 million payment. So why would banks elect to repay the investors of failed products despite having no legal obligation to do so?

“Banks are actually, formally or informally, on the hook for a lot of things not reflected on their books,” says Chovanec, reinforcing the point that the shadow banking system enabled by wealth management products is, in fact, central to China’s present financial system.

This fact of being informally on the hook, along with the size and structure of the system, poses what may be enormous risks.

The most immediate problem is liquidity risk, in other words a sudden shortage of cash, which stems from that switch beginning in 2012 from using WMPs for money market funds, to assets like property investments.

“One of the key problems with wealth management products is that there is quite a large asset/liability mismatch,” says Yuan of Fitch, echoing Wright’s explanation of how these products transformed. “You are financing medium to long-term assets with very short-term funding.”

Because WMPs typically mature in a matter months, funding for big projects needs to be continually rolled over by issuing new products to cover old ones and keep the cash flow going. This creates a tricky situation of dependence on the continued sale of WMPs. Should the investing public become spooked, say by a big single default, the money tap could turn off.

Not only are they heavily used by most banks, but they are particularly relied upon by mid-tier state banks, those that lack the enormous deposit bases of China’s largest state banks. Some smaller banks are issuing far more WMPs at this point than they are attracting deposits.

“A withdrawal of WMP funding [at mid-tier banks] could cause them to have a major liquidity crunch,” says Yuan. “If a number of them go to the interbank market at once… they may have to go in the last instance to the central bank.”

“It could get quite messy,” he adds.

This is partly why institutions have been willing to cover losses, to quell any investor jitters before they could get out of hand. According to the Financial Times, when Hebei Financing went bust, 11 shadow banks wrote an open letter to the Hebei government asking for a bailout, lest a bevy of WMPs default, including the ones that people protested over. The letter appealed to concerns about social stability.

There are a number of other amplifying factors on top of this, most notably the cross-ownership of WMPs between banks, which amounts to 15% of WMPs, according to official data. Such a practice creates the possibility of a default chain reaction, but deleveraging the situation is made devilishly difficult by the very nature of the system.

In the short term, any attempt to tighten rules on WMPs could itself cause a liquidity crunch. But the deeper concern may not actually be WMPs but the apparent need to keep economic growth running at a smooth clip.

“There’s a consequence to reigning it in, which is you don’t hit the GDP target,” says Chovanec. “So what happens? You don’t reign it in.”

The China Banking Regulatory Commission issued new draft rules in November aimed at curbing WMPs, specifically the repackaging of high-risk loans. But according to The Wall Street Journal, the broad language still gives banks quite a bit of leeway in interpreting the requirements, meaning they may be ineffective.

“China wants to have a correction without having a correction,” Chovanec says.

The bottom line is that investors believe that the government will not allow anything to fail, at least not in a widespread fashion, the net effect of which is to throw off the rational process of investment.

“When the system is structured in this way, there is a tendency for asset misallocation because… losses will be socialized,” says Yuan.

Put another way, it means the inherent potential of a given investment to generate returns is less important than the good faith of the backer—in this case the government.

Wright holds a similar view: “WMPs are a particular evolution in response to a system in which things aren’t assumed to fail. That’s the central part of this story.”

As Wright, Chovanec and Suslov all see it, it is a moral hazard. And it may be the case that it is an unavoidable one, in the context of the Chinese economy.

“If WMPs didn’t exist, someone would have to create them,” says Wright.

That begs the question: What then can be done?

Perhaps the most obvious fix would be to move WMPs out of the dark, and the Chinese government is already working to do this. In October the PBOC began requiring commercial banks to count them as part of their overall credit, according to The Wall Street Journal.

The second thing is clarifying risks to the buyers of WMPs, probably by making them more transparent.

“What is especially important here is making it really clear what the underlying assets are,” says Suslov. “Often times investors don’t understand, and often times the assets can be really complex and complicated.”

Next would be targeting the use of the WMPs as instruments, particularly because of the asset/liability mismatch. According to Wright, the entire world is moving toward longer-term, more stable bank financing, while China has done the opposite with short-term WMPs.

Really cracking down on them could be painful.

“It would cause widespread disruption,” says Wright, especially to the smaller banks relying upon them. “But because they are so short term in nature…you would pretty rapidly roll off 70-80% of WMPs in 3-6 months.”

However, Wright also notes that WMP growth is already beginning to slow down on its own, because good assets are drying up.

But the longer term implications are more fundamental to the way the Chinese economy operates, that is, the current reliance on credit for growth, and the fact that everything is at least implicitly guaranteed by the government.

As Chovanec puts it, by far the worst effect of “trying to run an economy with no losers” is that the market signals about what to really invest in don’t get sent, which pushes investment into things like real estate that keep the economy growing at least on the surface. As he sees it, at some point losses have to be accepted.

But the good news is there is still ample opportunity out there in the real economy.

“I think there are lots of areas in China’s economy that can generate real productivity gains, and you don’t even have to invent anything,” Chovanec says. “But the signals have to be [there].”