Modern history


China’s Captive Bond Market

“Compared with other financial instruments and against the backdrop of a high savings rate and high ratio of M2 to GDP, China’s corporate bond market has been developing very slowly and its role in economic growth has been rather limited. Such lack of development has also distorted the financing structure and produced considerable implicit risks, whose consequences may be grave for social and economic development.”

Zhou Xiaochuan

Speech at China Bond Market Development Summit

October 20, 2005

The demand from corporations and other issuers for cheaper capital than banks were willing or able to provide gave rise to the debt-capital markets in the developed economies. The basic assumption of issuers is that banks do not have a monopoly on understanding and valuing risk; large institutional investors, such as insurance companies and pension funds, also have the capacity to make investment judgments independently. So why rely only on banks for capital if you can get money more cheaply from other investors? Why not use markets to press the banks for cheaper funds? In China, over the past several years, a similar process appears to be happening. Its bond markets have enjoyed record issuance volumes, developed standardized underwriting procedures and allowed some foreign participation. Is it possible that in the not-distant future, investors in this market will compete with banks for corporate issuers and so take some of the credit- and market-risk burden from them, as has been one of the explicit reform objectives of the central bank?

In China, nothing is as it appears; words similar to those used internationally can have different meanings. Here, the markets were created by the same group of reformers who promoted bank reform. Beginning in 2005, with the aim of reducing excessive risk concentration in the banking system, they took over the largely moribund inter-bank market for government debt and introduced products modeled after those available to corporations internationally. On the surface, their efforts appear to have paid off. But huge issuance volumes, thousands of market participants and a growing product range do not alter the fact that China’s debt markets remain at a very primitive stage—an assessment with which no market participant in China would disagree, as Zhou Xiaochuan’s comments above attest. China’s debt markets are captive both to a controlled interest-rate framework on the one hand, and, on the other, to investors that, in the end, are predominantly banks. To understand why China’s bond markets are moribund requires digging into the technical details. But seeing how these markets are controlled is a key part of understanding how the Party manages China’s financial system: the symbols of a modern market are there, but the market itself is not.

Normally, the word “primitive” is used to indicate that the necessary market infrastructure is missing, but in China, all such infrastructure is in place. Like highways, new airport terminals or CCTV’s ultra-modern office building in Beijing, it exists because the Party believes bond markets are a necessary symbol of economic modernity. So there are ratings agencies (five), regulators (at least seven) and industry associations (at least two) with overlapping authorities and little respect for one another. There are now many of the same products that can be found in more developed markets, including government bonds, commercial paper (CP), medium-term notes (MTN), corporate bonds, bank-subordinated and straight debt, some asset-backed securities, and so on. These products are traded for cash, repo-ed1 out, or sold forward, and interest risk is hedged through swaps: all as might be expected.

What makes China’s bond markets “primitive,” however, is their lack of the engine that drives all major international markets. That engine is risk and the market’s ability to measure and price different levels of it. Risk, in market terms, means price; like everything else, capital has a price attached to it. In China, however, the Party has made sure that it alone, and not a market-driven yield curve, provides the definitive measure of risk-free cost of capital and this measure is based ultimately on the funding cost for bank loans, the one-year deposit rate. Consequently, in the primary (issuing) market for corporate debt, it is common practice that underwriting fees and bond prices are set with reference to bank loans, and not to true demand. Artificially low prices are then compensated for by the issuer’s agreement to an exchange of additional value outside of the market through, for example, conducting a certain amount of foreign-exchange transactions. In other words, bond-price setting is bundled with other business not in the market and the underwriter then holds the bond to maturity. Why? In the secondary (trading) market, investor demand is free to price capital, but the low issuing prices in the primary market mean that the bond underwriter will take a loss if he sells. Thus, the number of Chinese government bonds (CGB) and other bonds traded daily is in the hundreds at best. To the extent that bonds change hands, they do so at prices reflecting the premium that holders must pay to buyers to unload the security. If there is no active trading, there can be no accurate market pricing standards, only a price that might be called a “liquidity premium.”

There is an additional, historical, reason explaining the weakness of China’s bond markets. China is a country where the state—that is, the Party—owns everything and there is no tradition of private property. It might be expected, therefore, that the debt markets would have grown into the most developed market for capital. Unlike stock, debt does not touch directly on the sensitive issue of ownership. As even the most casual observer cannot help but note, however, everyone in China—from retail mom-and-pop investors to provincial governors and Communist Party leaders—is infatuated with stock markets. This has been true since the early 1980s when shares were “discovered” and is one of the main explanations for why observers believe that China is evolving along the path traced by developed economies.

So why not debt? The reason is simple: the government and SOE bosses quickly figured out that stock markets provide enterprises with “free” capital in the sense that it need not be repaid. In contrast, like a loan, bond principal must be paid back at some point and in the past, this often has proved to be “inconvenient.” Even better, a public listing provides an SOE with a “modern” corporate veneer (plus higher compensation levels for senior staff if the company is listed overseas) that issuing debt does not. Again, it’s the great attraction of symbols. Selling shares is a game-changer in these and many other ways, while issuing more debt is just business as usual. No Chinese CEO was ever lauded in the financial press for borrowing money from a bank.

China’s beautiful market infrastructure is necessary, but insufficient to raise the bond market above its primitive stage. As a result of manipulated pricing, corporate issuers are indifferent to the choice of debt instruments; bonds or loans are the same to them. More importantly, underwriters and investors are also indifferent to this market because they cannot make money. This chapter explains why this is so. Caught up in guidelines left over from the Soviet central-planning era, interest rates do not reflect true market forces, so debt valuations are distorted. But this is how the “system” likes it; the Party’s urge is to control. Party leaders believe they are better positioned than any market to value and price risk. The near-collapse of the international banking system in 2008 has only confirmed them in this belief.

What does bond market “development” mean, however, if not establishing over time a finely tuned understanding of the price of risk? Part of the notion of risk is that of change. But China’s debt-capital markets have from their inception been founded on the expectation that there will be no change, whether in the quality of issuer or in supply and demand as understood by developed markets. Zhou Xiaochuan’s remarks, therefore, are an almost-unique public indication that at least some senior officials are aware of the real systemic dangers being created by this suppression of risk. Given his expertise, his remarks on the consequences for social and economic development are not entirely surprising. If all this is true—that the market is creating risk—why does China need a bond market or, in any case, the one it has currently?


The fact that banks hold over 70 percent of all bonds highlights the importance of this question. For the group of market reformers surrounding Zhu Rongji and Zhou Xiaochuan, developing the bond market was a basic part of the bank-reform process that began in 1998. A strong bond market would encourage institutions other than banks to hold corporate debt, and risk could be diversified. But if the markets are not wholly opened to the participation of investors not controlled by the state, this cannot happen. The reality is that China’s bond markets has evolved over the past 30 years because the national budget needs to be financed; however, its tax-collecting capacity was, and remains, too weak. If corporate investors could rely on bank lending, the MOF cannot, not if it follows in the model of state treasuries elsewhere in the world. What would a minister of finance be if he could not issue government bonds? What would a modern economy be if it didn’t have a government-bond yield curve to measure risk? The MOF’s growing demand for funds from the early 1980s led to the creation of a narrow market, which reformers 20 years later would seek to broaden.

In the early 1980s, markets for securities of any kind did not exist in China. The last bond the country had issued was in 1959 and all knowledge associated with it had long since been lost to the Cultural Revolution. But ambitious national budgets in the early 1980s began to create small deficits (see Figure 4.1). Confronting the question of how to deal with increasing amounts of red ink, the idea of issuing bonds was voiced by a courageous person at the MOF. This raised questions about the identity of the investor base and what price to pay them. At the start, only SOEs had money (of course, all borrowed from the banks), so by default, they were compelled to fund the government budget as a political duty. As for price, bond interest was set administratively with reference to the one-year bank-deposit rate set by central bank fiat to which was added a small spread. As the data in Table 4.1 indicate, individuals received a higher interest rate than SOEs, which reflected both the MOF’s need for third-party funding as well as the demand of retail investors for a reasonable yield. This was a real market situation and the MOF had yet to find a way to minimize its funding costs. As for a secondary market for debt, there was none. SOE investors were forbidden to sell bonds based on logic relating to the MOF’s “face”: selling a bond was seen as a lack of confidence in the state’s creditworthiness.

TABLE 4.1 Composition of national savings and sales of bonds, 1978–1989

Source: Gao Jian: 47–9; China Statistical Yearbook, various

Note: All coupons for maturities of minimum 5 years


FIGURE 4.1 National budget deficits vs. MOF issuance, 1978–1991


Source: China Statistical Yearbook; Wang Nianyong, p. 53

Note: Includes central and local government budgets; excludes maturing bonds rolled over in 1989–1991.

Over the course of the 1980s, successful agricultural reforms and the growth of small enterprises in the cities rapidly enriched the general population. By 1988, nearly two-thirds of all bonds were sold directly to household investors. Then, from 1987, the real market turned as inflation boomed and banks were ordered to stop lending. SOEs and individuals, strapped for cash and seeing yields turning negative, discovered they could sell off their bond portfolios, although at deep discounts, to “speculators”. Suddenly a wholly unregulated over-the-counter (OTC) secondary market sprang into existence just as the craze for shares hit its peak in 1989 and 1990. Here were China’s first (and still only) true markets for equity and debt capital! They were rapidly closed down.

When the political dust from June 4 had settled, China in 1991 had the beginnings of regularized bond and stock markets, but they were ensconced safely inside the walls of the new Shanghai and Shenzhen exchanges. The new infrastructure suggested that market reformers had prevailed, but the truth is they had been forced to compromise away the heart of the markets. The two exchanges existed only to provide controlled trading environments where prices and investors could be managed to suit the government’s own interests. For its part, the MOF had also realized by this time that its fund-raising difficulties in part reflected investors’ fear of locking up their cash with no legal way to recover it until their bonds matured. To expand its own funding sources, therefore, from the early 1990s, the MOF began to develop a secondary market on the exchanges.

Proper pricing of the bonds remained a problem, however, and it wasn’t until 1994 that the MOF stumbled on a workable combination of underwriting structures and market-based bidding within the strictures of PBOC interest rates that allowed CGB issue amounts to increase (see Figure 4.2). The innovator at the MOF, Gao Jian, loves to recount the story of how he created a Dutch auction-based bidding system for a loose group of primary dealers using a Red Tower Mountain cigarette carton to hold their bids.2Someone’s smoking habit and an equitable way to distribute the obligation to underwrite government debt largely solved the MOF’s fund-raising difficulties and created the market infrastructure that could be used a decade later.

FIGURE 4.2 Debt issuance by issuer type, 1992–2008


Source: PBOC Financial Stability Report, various; China Bond

Note: The 2007 government bond number excludes the RMB1.55 trillion Special Treasury Bond used indirectly to capitalize China Investment Corporation.


In spite of the success of Gao’s cigarette carton and Dutch auctions, underwriting CGBs, as well as corporate and bank debt, has remained very much a political duty, just as it had been from the beginning. This can be seen from the simple fact that the market did not, and still does not, trade. What is a market without trading? The reason for the lack of liquidity is straightforward: bond prices in the primary market are set below levels that reflect actual demand. Despite its surfaces—record issuance, improved underwriting procedures and issuer disclosure, and even a grudging openness to foreign participation in some areas—it is less a market to raise new capital at competitive prices than a thinly disguised loan market.

This reality is highlighted by the fact that of the primary dealer group of 24 entities, all but two are banks.3 With the sole exception of the NDRC’s recondite enterprise bonds (qiyezhai image) that are underwritten by securities companies, banks are the dominant underwriters of all bonds including CGBs, PBOC notes and policy-bank bonds. They underwrite and hold the bonds in their investment accounts until maturity, just like loans. Due to the skewed pricing mechanism in the primary market, banks, like their cousins, the securities companies have not developed the skills to value capital at risk. They do not need to: the PBOC does it for them by fixing the official CGB trading prices in the market as well as the even-more-important one-year bank-deposit rate.

PBOC’s perfect yield curves

To fully appreciate why there is no “market” in China’s bond markets requires delving into the meaning and practice of bond “yield curves.” These curves show the relative level of interest rates payable on similar securities of different maturities (see Figure 4.3for examples) and “cost” means what investors demand for a given level of risk. The interest rates payable by government, or sovereign, issuers are used as the basis of bond-underwriting decisions in all developed markets. This is founded on the theory that countries do not go bankrupt (which is clearly disputable) and that, therefore, they represent the risk-free standard in a given nation’s domestic bond market. In China, the MOF represents the sovereign issuer, the highest credit possible, and the Chinese credit-ratings agencies place the MOF as a unique risk category a level above the AAA (Triple-A) rating represented by, for example, PetroChina. It sounds much better to be a “Quadruple A” than the “Triple A” of, for example, the US Department of Treasury, which one Chinese agency has cheekily assigned it in the Chinese system. Figure 4.3 shows the PBOC-mandated minimum credit spreads for a variety of enterprise-bond credit ratings over the cost to the MOF. These curves show an ideal world that does not exist: why?

FIGURE 4.3 Mandated minimum spreads over MOF by tenor and credit rating


Source: China Bond, as of October 20, 2009

As in other international markets, the curves are based on the underlying MOF yield curve; for example, the minimum 10-year AAA-to-MOF spread is circled.4 The trouble in China, however, is that the MOF yield curve is disregarded in favor of the PBOC’s bank interest rates on loans. It is disregarded because it does not truly exist, as is explained below. The PBOC-mandated one-year loan and deposit rates used by banks are shown in Figure 4.4.

FIGURE 4.4 One-year PBOC RMB bank deposit vs. lending rates, 2002–2008


Source: China Bond

The regulated spread between the cost of funds on deposits and minimum bank lending rates is deliberately set to provide lenders a minimum guaranteed 300 basis point (three percent) profit.5 When a bank underwrites a bond, therefore, it will, among other things, compare its potential return with that of a loan of a similar maturity to a similar borrower. The issuing company will, of course, consider the same thing. To the extent that this comparison to loan rates influences the underwriting decision, bond pricing does not reliably reference the MOF yield curve. In actual practice, the MOF curve is frequently disregarded and corporate and financial bonds are priced lower than the curve would indicate. This is because banks are motivated by compensation from the issuer via other supplemental businesses. But they also know full well that, as mentioned, the MOF yield curve is a fiction.

Fictional curves from fictional trading

Figure 4.5 shows an actual picture of a single day’s corporate-bond trading on December 8, 2009. The yield curves presented look like something created by random machine-gun bursts against a wall. What the figure represents can be understood by examining the two highlighted AAA transactions, both for bonds with tenors of around five years. As can be seen, the trades were completed at wildly different levels—a low of less than two percent and a high of close to five percent. These are not unique transactions; the chart abounds in such examples. Why was there such a differential between two seemingly similar securities?

FIGURE 4.5 Actual enterprise-bond yield-curve data, by credit rating


Source: Wind Information, December 7, 2009; excludes MOF, CDB, and financial bonds

The absence of active market trading explains this strange data. On December 8, 2009, for example, the entire China inter-bank market for corporate bonds recorded only 1,550 trades—this in a market comprising over 9,000 members and RMB1.3 trillion (US$190 billion) in bond value. In contrast, the US Treasury market each day averages 600,000 trades comprising US$565 billion in value. If market participants do not actively trade, how can the price of a bond be determined and serve as a meaningful measure of value?

The true character of China’s bond market becomes even clearer when the focus is put on only MOF and CDB bond trades, as shown in Figure 4.6. On December 7, 2009, the number of trades totaled 52 for MOF bonds and 108 for CDB. These numbers could, in all likelihood, be halved since market-makers create volume (as they are required to do) by, among other actions, selling a bond to a counterparty in the morning and buying it back in the afternoon. With trading volume so light, whatever yield curves that can be drawn are almost arbitrary. How then can the MOF curve be considered a meaningful pricing benchmark for corporate-debt underwriters or, indeed, corporate treasurers?

FIGURE 4.6 MOF and China Development Bank bond trades


Source: Wind Information, December 8, 2009

Fixing a yield curve

The data raise a question regarding the basic quality of the MOF curve represented in Figure 4.3. China uses what are called in the financial industry “daily price fixings” for its debt securities. This means that there is an “official price” set for traded products such as foreign exchange or securities. Usually, this is done by the central bank or market regulator in consultation with a number of market participants and is necessary because the given product did not trade or traded too few times for the market to establish a price.

Official price setting is not an uncommon practice: there are fixings even for such actively traded products as the Japanese yen, as well as such partially convertible currencies as the Indian rupee and, of course, the Chinese renminbi. In China, since October 2007, this official “fixing” for bonds has been done by the China Government Securities Depository Trust and Clearing Corporation, a nominally “independent” entity owned by the PBOC and the depository for all bonds traded in the inter-bank market. Bloomberg carries the China Bond daily fixing table for CGBs and CDB bonds, as illustrated in Table 4.2. The actual traded price information for each bond traded on that day is also shown.

TABLE 4.2 China Bond price-fixing data, January 4, 2010

Source: Bloomberg, China Bond, and Wind Information. All fixed-rate bonds


The trading data show that on January 4, 2010, there were a total of 32 CGB trades with a combined value RMB5.57 billion and 55 CDB trades with a combined value of RMB29.53 billion. The actual trades used in the daily fixing table were extracted from this voluminous activity and illustrate precisely why Chinese sovereign yield curves are more fiction than fact. For the one-year to five-year section of the CGB yield curve, not one of these bonds traded, not even once! This official yield curve, built out of nothing but assumptions, did dovetail nicely, however, with the six-, seven- and 10-year bonds, which traded a grand total of five times that day.

Data for the MOF and CDB bonds shown in Table 4.2 have been charted in Figure 4.7. Together, they describe a smooth, upward-sloping yield curve. Against this background, what reliance should market participants put on CGB yield curves or, in the case of the CDB bonds, to a notional spread over treasuries? It is not surprising, therefore, that the absence of trading for what should be the most liquid products characterizes the market as a whole as well.

FIGURE 4.7 MOF and CDB “fixed” yield curves, January 4, 2010


Source: Wind Information, January 4, 2010

Also on January 4, the entire inter-bank market saw only 615 trades (see Table 4.3), among which CGBs incredibly traded the least of all and represented only 3.3 percent of the total value traded. In contrast to the US$25 billion in bond value traded that day in China, the average daily trading volume in the US debt markets is US$565 billion, a figure itself far in excess of the average total daily global equity trading of US$420 billion.6 These US trades result on average in over US$1 trillion in bond value moving between accounts each day on the US Federal Reserve Bank’s electronic settlement system.

TABLE 4.3 Inter-bank bond trading summary, January 4, 2010

Source: Wind Information


If the price points, “fixed” or not, on Figures 4.5 and 4.6 measure anything, it is the liquidity premium paid by investors to sell their bonds into a saturated market. This accounts for the widely scattered price points around faint yield curves. The story is summed up in Figure 4.8, which illustrates the fundamental illiquidity of CGBs, corporate and financial bonds. In all of 2008, for example, MOF bond turnover was only RMB3.5 trillion or about 10 percent of all outstanding CGBs. The most liquid securities are the shortest in tenor—MTNs, CP and PBOC notes—but even these turned over less than one time during the entire year.

FIGURE 4.8 Inter-bank market trading volume and turnover, FY2008


Source: PBOC

In sum, the absence of active market trading limits the price-discovery function of China’s bond markets. In turn, unreliable prices mean that the market participants cannot value risk accurately. A simple question such as how much a AA issuer would have to pay investors to buy its 10-year bonds cannot be answered with any certainty. On the other hand, China’s market investors don’t really care. Why should they when the majority of bonds offer “riskless” yields well over the one-year bank-deposit rate of 2.25 percent but, at the same time, well under demand in the secondary market? As long as inflation remains under control, why shouldn’t banks be happy to hold the bulk of these securities to maturity, just as they do their loan portfolios?

Cash vs. repo markets

China’s repo markets illustrate just what liquidity means in a bond market. Figure 4.9 shows the seven-day repo interest rate for 2008. Contrast the active trading in interest rates here with the anemic yield curves traced by the CGBs and CDB bonds shown in Figure 4.6. Clearly the cost of capital is being driven by supply and demand. What accounts for such trading? The wildly speculative bidding on shares offered in Shanghai IPOs forces investors to put together the largest amount of funding possible to secure an allotment in the share lottery. In IPO subscription lotteries, massive amounts of capital—often equivalent to tens of billions of dollars—are frozen to secure allocations of shares. A large portion of these funds is raised by repo transactions. This market, however, is much more akin to the pure short-term inter-bank loan market than to the long-term capital-allocation function of bond markets. The point, however, is that demand drives the price of capital here, but not in the bond market.

FIGURE 4.9 Seven-day repo volumes, interest rates vs. capital frozen in IPO lotteries


Source: Wind Information

Note: “Offline Frozen” indicates amount of capital used in bids for shares in the institutional “offline” IPO lottery.

As is obvious to even infrequent observers of China’s economy, speculation is a fact of life. This largely stems from the artificially fixed returns on bank deposits, loans and bonds, the only available investment alternatives outside of real estate, shares and luxury goods. Set at levels unreflective of the true demand for capital, the managed rates for these products create a stillborn fixed-income market and force investors to speculate. Capital gains, which are untaxed, are the main play for investors in China, whether retail or institutional, and none can be found in the debt-capital markets.

The “327” Bond Futures Scandal

If any one incident highlights why the government seeks to strictly control markets, it must be the bond futures scandal of 1995. This story is already ancient history, but it explains why there is still no financial-future product of any kind in China’s capital markets.7 At its simplest, the scandal was a struggle between a major local broker backed by the Shanghai government, and the MOF; in other words, between local and central government interests. Wanguo Securities, owned by the Shanghai government, received inside information that the MOF planned to issue 50 percent more bonds in 1995 than it had the previous year. Expecting this larger volume to offset any gains from declining inflation, Wanguo’s traders, in contrast to the overall market view, expected bond prices to remain low. Over the early part of 1995, they accumulated a huge (and illegal) short position in bond futures contracts, in particular, the March 27 contract (which gave the scandal its name). News of this leaked out (nothing in China remains secret for long) and other market participants began to accumulate long positions, expecting prices to be higher in the future. This trend increased when other brokers learned that the MOF had determined to significantly reduce its issuance plans. Somehow, Wanguo remained ignorant of this and continued to build its short position in an effort to corner the market.

Acting through its wholly-owned China Economic Development Trust and Investment (China Development Trust; Zhongjingkai), the MOF took a corresponding long position. As the head of the China Development Trust was Zhu Fulin, the former Director of the Treasury Bonds Department of the MOF, this was never going to be a fair fight. When the MOF at last announced its much-reduced issuance plans and bond prices remained high, Wanguo frantically sought to square its position during the last eight minutes of market trading. Market volumes soared to unprecedented levels. By the end of the day, Wanguo’s actions had driven prices down but at the cost of a market collapse and the technical bankruptcy of many other brokerages. That evening, the Shanghai exchange, facing the reality that the futures market had collapsed, canceled all trades that had taken place in the last 10 minutes of trading and closed the market for three days so that contracts could be unwound and renegotiated. This meant that Wanguo itself faced being bankrupted.

An investigation ensued and Wanguo’s chairman, a respected founder of the Shanghai exchange, was arrested and later sentenced to 17 years in prison. The fallout continued when Wanguo itself was merged with Shenyin Securities, then Shanghai’s second-largest firm, to become today’s giant Shenyin Wanguo. The very reformist chairman of the CSRC, Liu Hongru, took responsibility, although he had no direct control over the exchange at this time, and the financial-futures product was eliminated and remains so. Soon thereafter, Beijing took over control of both securities exchanges. Shanghai was most definitely the loser in this battle.

In this zero-sum game, someone had to be the winner and, of course, it was the MOF. China Development Trust was rated the top broker on the Shanghai exchange for 1995 “due to its massive trades in treasury bond futures . . . accounting for 6.8 percent of total annual exchange turnover.”8 Politically astute, China Development Trust seems not to have booked for itself what must have been massive profits. Rather, it allowed its “clients,” who no doubt included the MOF, to do so. In the following years, this powerful company became a major institutional market manipulator, whose actions could be seen in some of the most outrageous cases of stock ramping and corporate collapse. However, given its MOF background, Zhongjingkai escaped censure and closure until Zhou Xiaochuan finally closed it in 2001. It was not the only such institutional player with a central government background.

Ironically, one month before the 327 Incident, Vice Premier Zhu Rongji, who was then responsible for the financial sector, had fiercely criticized the rampant speculation in the bond futures market “by a number of huge interest groups, taking funds of the state, the local governments and the enterprises to seek profits.” Zhu had identified a growing problem but was apparently unable to do anything about it. He could, however, eliminate the futures product. Given the political cost associated with this scandal, it should not be surprising that the Party prefers an orderly, controlled bond market, even if this is, after all, moribund. But by refusing to reform the market, the Party simply promotes the forces of speculation which, as China has become more prosperous, become all the stronger.


At the base of China’s bond and loan markets are China’s household savers. Today, banks hold more than 70 percent of all bonds in value terms, but this was not always the case. In the earliest days of the market in the 1980s, individuals became the dominant investors, annually snapping up 62 percent of all bond purchases. By 2009, however, they had nearly disappeared from the field, accounting for only one percent in outstanding bond value (see Figure 4.10). Foreign banks account for another seven percent, which means that state-controlled entities hold 92 percent of total bond investment. What’s more, many of these same state entities are the only issuers in the market.9

FIGURE 4.10 Change in types of bond investors, 1988 and 2009


Source: 1988, Gao Jian: 49–51; 2009, China Bond

This fact has profound implications for China’s financial system. If the markets today simply function as clearing houses that move money from one pocket of the state to another, then they have developed away from their more diverse origins in the 1980s into something resembling a pyramid scheme. This is exactly why Zhou Xiaochuan has described them as “distorted” and filled with “implicit risk.” Why has the role of the critically important non-state investor become so diminished?

As part of its effort to develop greater market capacity, in 1991 and 1992, the MOF experimented with different underwriting methods. Its own experience had clearly highlighted the problems limiting large-scale bond issuance. First of all, there was the pricing problem. But, secondly, the over-reliance on the retail market created major difficulties. As individuals purchased bonds in small amounts, simple logistics limited the total amount of bond issuance and offering periods were often up to six months before an issue could be closed. Even to access these investors, the MOF found itself having to pay close to market prices. The retail market also tended to buy and hold until maturity, thus inhibiting the emergence of a secondary market. Finally, maturities tended to be short as a result of both inflation and retail preference. Small issue sizes, high cost, shorter maturities and the fact that there was no secondary market prevented the development of benchmark interest rates and, ultimately, meaningful yield curves. All of these are legitimate reasons to seek to develop an institutional investor base.

The MOF had sought early on to develop institutional investors by seeking support from banks and non-bank financial institutions. However, banks in the 1980s had little excess liquidity and, therefore, little capacity to invest. Even if the State Council had allowed the MOF to develop a market-based pricing method, the retail nature of the investor base may have limited its ability to raise funds in line with its needs. It was at this point that the story of stock markets and bond markets came together. Having created stock exchanges to manage the “social unrest” associated with street trading, the government also brought bonds “inside the walls,” especially those of the Shanghai Stock Exchange.

The exchanges enabled demand to be sourced from both individual and institutional investors; all were members of the new markets. The banks also had much deeper pockets given rapidly growing retail deposits (see Table 4.4) and it was not long before the government began to lean on them for support as they discovered an interesting fact.

TABLE 4.4 Composition of Big 4 bank deposits, 1978–2005

Source: China Financial Statistics 1949–2005


Accessing funds from the banks had the effect of lowering the MOF’s interest expense. The Party could urge banks to buy bonds at levels just above the one-year rate they were paying retail depositors, while retail investors using the same bank deposits to buy bonds would require far higher returns. In other words, the banks provided the government with direct access to household deposits at government-imposed interest rates without even having to ask the depositor for permission: the banks simply disintermediated them. Unlike unruly retail investors seeking to maximize returns, banks had the pleasing aspect that their senior management (Party members) did as they were told. The Party was now easily able to direct funds where it wanted and in the amount it wanted without the need for excessive cajoling or paying market rates. Meanwhile, it could persuade itself that this was the right thing to do since it “protected” the household depositor from undue credit risk.

At first, there was no conflict of interest: individuals were crazy about shares, not bonds, and banks could not buy shares. But as China emerged from the major inflation of the mid-1990s, bonds suddenly offered a very attractive return in comparison to a collapsing stock index. The problem was that retail investors were unable to get their hands on them. In just a brief period of time, China’s banks had monopolized bond trading on the Shanghai exchange. The story goes that a feisty Shanghainese housewife complained about this de facto government-bond monopoly and her anger reached all the way to Zhu Rongji. Characteristically, Zhu took decisive action and in June 1997, he summarily kicked the banks and the bulk of government bond issuing and trading out of the exchanges and into what was then a small and inactive inter-bank market.10 Since then, individual investors have been limited to buying savings bonds through the retail bank networks and institutional investors have been largely limited to the inter-bank markets.11

This significant structural change meant that although the market continued to rely overwhelmingly on the state-owned banks, all other state-owned entities that could qualify as members could also participate (see Table 4.5).

TABLE 4.5 Number of investors, October 31, 2009

Source: China Bond

Note: Members include individual branches in the case of institutions.

Member type


Special members (PBOC, and other government agencies)


Commercial banks


Credit cooperatives


Non-bank financial institutions


Securities companies


Insurance organizations




Non-financial organizations


Total inter-bank market members


Individuals (not members of the inter-bank market)


In short, bonds returned to their earliest stage, when the state was its own investor. But the principal difference was that banks and all other non-bank financial institutions replaced SOEs, which meant that household savings would be channeled directly by the Party. This explains how the banks came to hold over 70 percent of all fixed-income securities in China, including 50 percent of all CGBs, 70 percent of policy-bank bonds, and nearly 50 percent of commercial paper and medium-term notes issued (see Figure 4.11). Only in the case of the NDRC’s enterprise bonds do insurance companies (NBFI) displace the banks as the principal investors, holding some 46 percent of these securities.

FIGURE 4.11 Investor holdings of debt securities, by issuer, October 31, 2009


Source: China Bond

Note: Non-state investors include foreign banks, mutual funds, and individuals.

In the international markets, banks also dominate underwriting and trading, but investors and their beneficial owners are, of course, far more diverse, with large roles being played by mutual and pension funds as well as insurance companies. In China, such diversity is beside the point since all institutional investors, whether banks or non-banks, are controlled by the state. In such circumstances credit and market risk cannot be diversified. This is why China’s markets remain primitive and why there is the “implicit risk” alluded to by Zhou Xiaochuan.

In late 2009, the CBRC suddenly became aware of this inevitable reality when it stopped all issuance of bank-subordinated debt. Why had it been oblivious to this risk from the beginning? If the state owns China’s major banks outright, as it does, what is the significance of Bank of China issuing subordinated debt to investors that are largely other state banks? The state is simply fooling itself by subordinating its own capital to its own capital. The level of risk in the system has not changed one bit, even if the financial landscape seems the richer for adding this new product.

All of this raises the question of why it has been so difficult for foreign banks and other financial institutions to become involved in this market. Over the past 15 years, China’s leaders have witnessed the Mexican debt crisis of 1994, Argentina’s peso crisis of 1999 and the ongoing sovereign-debt crises of Greece and Spain. They have seen the huge ramp-up of their own stock index in 2007 and its collapse in 2008. Local newspapers and other media commentary are rife with talk of hedge funds, hot money and unscrupulous investment bankers. An inherently conservative political class, whose natural instinct is to control, will not easily invite those it cannot easily control to participate actively in its domestic debt markets. But, as appearances have to be preserved, there will always be slight movements toward market opening. But there will be no true opening.

What would happen to bank and insurance company holdings of CGBs or other corporate and financial bonds in an inflationary environment? As mentioned, China’s central bank manages interest rates in order to contain change because change is risk. No matter that these state institutions hold fixed-income securities as long-term investments to avoid marking their value to market, in an inflationary environment their value will inevitably decrease as funding costs rise. The inevitable result would be a growing drag on bank income even if valuation reserves are not taken. This problem can be seen clearly in bank financial statements. For example, the auditors for ICBC’s 2009 financial statements usefully show separately the yields on the bank’s loan, investment-bond, and restructuring-bond portfolios (see Table 4.6).

TABLE 4.6 Yields on loans, investment and restructuring bonds, 2008–2009

Source: Bank FY2008 financial statements

Note: * CCB and BOC bond rates are calculated on portfolios that include the restructuring securities; hence returns are pulled down. ICBC rates have been separately calculated.


The restructuring bonds yield on average almost exactly the one-year bank-deposit rate and are fixed. In other words, in a low-inflation environment, they will nearly break even, whereas the bonds held as investments yield 3.34 percent, around 1.1 percent over the one-year deposit rate. This is somewhat better, but raises the question: why hold such huge bonds portfolios when loans yield on average nearly seven percent? Banks hold these portfolios partly for liquidity reasons, but largely because they are required to do so by the Party. If the ultimate objective of bank management were to maximize profit, would such low-yielding bonds make up 20–30 percent of their total assets (see Figure 4.12)?

FIGURE 4.12 Composition of total assets of Big 4 banks, FY2009


Source: Bank 2009 annual reports

Interest-rate risk holds true for all bonds, but corporate bonds also have a credit aspect. In the event of their inability to pay interest, banks will experience a drag on income and, sooner or later, would be compelled to re-categorize their internal credit ratings and make provisions as the bond becomes, in effect, a problem loan. Even if the bank could sell the bond into the market under such circumstances, it would be forced to take an outright loss. China’s major financial institutions, banks and insurance companies are all listed on overseas exchanges and audited by international firms. The need to take reserves should be unavoidable in such circumstances. China has not been, and will not be, exempt from such circumstances.

In short, China’s banks face severe challenges on three fronts. In addition to their structural exposures to the old NPL portfolios of the 1990s, there will inevitably be new NPLs arising out of their lending spree of 2009. Thirdly, the banks are fully exposed to both interest rate-related and credit-induced write-downs in the value of their fixed-income securities portfolios. In 2009, securities investments constituted 30 percent of the total assets of China’s Big 4 banks, or RMB7.2 trillion. While the interest risk of these portfolios can now be hedged somewhat as a result of the very recent emergence of a local-currency interest-rate swap market, for the state, it is a zero-sum game: BOC may effectively hedge, but its counterparty will almost inevitably be another state-owned bank. The effect of mark-downs, credit losses or even simply negative yields on bank capital would obviously be significant. From the viewpoint of the issuer, too, they seem to make little difference. In the international markets, corporations can source cheaper funds from other classes of investor; but in China, the banks remain the investor and the all-in cost to the issuer will be the same as a loan. So the question again presents itself: why did China build its fixed-income market?


1 A “repo” or “repossession” contract is a kind of financing transaction in which a party holding, most commonly, government bonds provides the bond as collateral to a second party who then lends money to the first party. This is a cheap way of funding a large bond portfolio.

2 For the only authoritative history of China’s government bond markets, see Jian Gao 2007.

3 This group is not the same as the primary-dealer group authorized by the MOF for CGB underwriting. The two non-banks are CITIC Securities and China International Capital Corporation. In late 2009, some foreign banks received licenses to underwrite financial bonds only to be told that “circumstances are not yet mature” for their active participation.

4 Of course, underwriters are free to set higher interest rates (known as coupons) on enterprise bonds if their issuer clients agree.

5 From January 2004, the cap on maximum interest rates on loans was eliminated, but banks are still subject to a minimum rate charged, which is 90 percent of the PBOC set rate for the relevant tenor.

6 Figures from US Department of Treasury, Office of Debt Management, June 2008.

7 There is, however, now a stock index future product.

8 Foo Choy Peng, “China Economic Rated Top Broker,” Bloomberg, January 13, 1996.

9 The Asian Development Bank and the International Financial Corporation, a part of the World Bank, have been the only foreign issuers in the domestic bond market to date.

10 The inter-bank market in China was established in 1986 as a funding mechanism for banks in which those with surplus funds place them with others needing additional funding in order to balance their books.

11 A small number of bonds remained listed on the Shanghai exchange to enable securities firms to finance themselves through repo transactions. Until recently, banks were excluded from this market. Their reintroduction is largely an effort to merge what has become two separate markets: the exchange-based and the inter-bank.

If you find an error or have any questions, please email us at Thank you!