A Primer on China's Financial Data Quality

Chinese data is bad.

Terrible actually, when compared to the financial data standards of the top global economic powers. That statement is a matter of empirical observation, which we will explore later, but it also reflects a consensus and Narrative surrounding the ‘China Story.’ Like all capital-N Narratives, while this one did originate in truth, it now oversimplifies the matter and obscures subtleties which should be important to all strains of China watchers.

I’d like to dig deeper into this Narrative to unearth a more nuanced understanding of Chinese financial data. These are the interesting questions: Why is Chinese data bad? What does bad data look like? And, what can I do about it?

Answering those questions requires a grounding in the institutional arrangements that produce the incentives which govern Chinese market participant behavior. But first, let’s begin with the story of one of China’s more ignominious forays into Western capital markets.

Table of Contents:

  1. Introduction: From the ‘China Growth Story’ to the ‘China Data Story’
  2. Institutional Arrangement I: State Ownership of Listed Firms
  3. Institutional Arrangement II: Government Control of Capital Markets
  4. Institutional Arrangement III: Weak Legal Institutions
  5. Prospects for Improvement: Partly Cloudy
  6. A Bright Spot: Chinese Analysts
  7. Conclusion: Where Do I Go from Here?

Introduction: From the ‘China Growth Story’ to the ‘China Data Story’

Sino-Forest, one of China’s leading commercial forest plantation operators, listed on the Toronto stock exchange in 1996 via reverse takeover. Benefiting from years of double digit economic growth driven by infrastructure and real estate construction, Sino-Forest rode the wave of the China growth narrative to garner a $7 billion valuation. The company was a slam dunk. Canaccord Genuity, a Canadian Investment bank, even went on to gush in a report that, “We continue to rate Sino-Forest as one of our top picks due to its strong long-term fundamentals based on growth of Chinese consumption.”

Then, in 2011 a Muddy Waters report accused Sino-Forest of extensive wrongdoing including, but not limited to, falsely overstating timber assets by $800 million and operating a complex Ponzi scheme to inflate earnings. The next month, the Ontario Securities Commission suspended Sino-Forest shares from trading. Since then, the firm filed for bankruptcy and investors recovered over $150 million in damages from a class-action lawsuit.

Sino-Forest was a fraud.

Much like Sino-Forest, the narrative surrounding China flipped quickly. China was a tremendous “growth story,” until, of course, it wasn’t. Now, the consensus position on investments in China is concern over the poor quality of financial and economic data. It’s become commonplace for the international financial media, such as The Economist, when reporting financial data to append editorializations like, “The problem, a familiar one in China, is that no one much believes these figures.”

So why is Chinese data so bad? Let’s begin our discussion with this summary given by the Stanford accounting professor Joseph Piostrowski and dean of the Hong Kong Business school, T.J. Wong:

The weak financial reporting and governance practices of Chinese listed firms can be viewed as an equilibrium outcome driven by China’s social, legal, and political institutions.

Allow me to submit a translation into lay (non-academic) English: The poor quality Chinese financial data is not the result of the unique mendacity of Party cadres or local management, nor is it the result of the general level of “development.” China’s Data quality is an “equilibrium outcome” – that is, no one has an incentive to change his or her behavior – given the local institutional arrangements, which are quite unique. In other words, incentives depend on institutions.

But before exploring what makes China’s institutional arrangements unique, let’s first understand the incentives that would motivate market participants to produce, disclose, and demand high quality financial information (HFQI).

An extensive literature on Western capital markets shows that firms with very transparent disclosure practices gain economic benefits such as a reduced cost of capital, higher valuations, greater foreign participation. Shareholders also benefit from a better allocation of resources by company management. On the flip side, firms may also stand to benefit from opacity. For example, majority shareholders can hide their use of control rights to capture private benefits at the expense of minority shareholders, or management can suppress information to disguise rent-seeking activities.

In China, the balance of incentives tilts in favor of opacity. We could also say that in China there just isn’t a strong demand for HQFI.

If you’ve read this far then you have dutifully accepted the premise that China’s data is in fact bad; but what exactly does that mean? Essentially, ‘bad data’ means that investors and lenders lack the information necessary to differentiate between a good company and a bad one, which is why so many Western investors completely avoid the country. In that type of environment, stocks are essentially commodities.

The measure of commodification used by academics is called stock price synchronization; it measures the percentage of stocks that exhibit price co-movement each week. For example, if all stock prices went up together this week, then the synchronization would be 100%. If half of the stocks went up and the other half went down, then the synchronization would be 50%. In a market where investors struggle to differentiate between good and bad companies, we would expect to see a very high level of stock price synchronization.

Over the last 15 years Mainland-listed firms have posted significantly higher synchronicities than those trading on the New York Stock Exchange, clearly illustrating the commodification of Chinese equities – the result of, at least partially, bad data.

Without further preamble, let’s dig into the unique institutional arrangements in China that disincentivize market participants from producing, disclosing, and demanding HQFI. They are,

  • State ownership of listed firms.
  • Government control of capital markets.
  • Limited protection of property rights.

Institutional Arrangement I: State Ownership of Listed Firms

China’s most interesting contradiction is how a communist country facilitates the operation of some of the world’s largest stock markets. Most analysis of what makes China unique begins here. For our purposes, we will examine how the state ownership of listed firms negatively impacts the information environment in China by disincentivizing market participants from producing and disclosing HQFI.

The first mechanism curbing demand for HQFI in China is the high ownership concentration of listed SOEs. For example, 86.5% of listed shares of PetroChina – China’s largest oil producer – are owned by the PetroChina group, a holding company which is itself wholly owned by the State-owned Assets Supervision and Administration Commission (SASAC). 

This highly concentrated ownership structure allows the state to directly monitor the performance of listed SOEs through management appointments made by SASAC and the Organization Department of the Communist Party of China. In this environment, where managers face powerful political incentives to toe the line, shareholders (the state) depend comparatively less on the monitoring function of accounting disclosures and more on internal political networks. A symptom of this – as well as other institutional arrangements that we will get to later – is that hiring high quality audit firms, such as the Big Four, is an unnecessary expense that few firms in China choose to incur.

The second institutional arrangement perverting data disclosure incentives for SOEs is the absence of bankruptcy risk.

The government can and does bail out troubled firms with cash subsidies, preferential bank loans, or politically directed transactions. So, firms gain little in terms of either a reduced cost of capital or a higher valuation from improving transparency. Additionally, institutional investors in these firms, who are normally producers of firm-specific value-relevant information, enjoy an implicit downside protection which limits their incentive to dig deeply into the financial performance of the SOEs – they are only interested in seeing which firms boast the best political connections.

Third, SOE managers are responsible not just for maximizing profit, but also for maintaining high levels of employment and advancing national policy objectives. Thus, shareholders of SOEs – the state – rely on metrics outside of accounting disclosures which further reduces the demand for HQFI.

Fourth, many SOEs adopt China’s unique corporate group structure involving a single upstream owner of the listed entity and multiple layers downstream of subsidiary firms. (Note, for more on China’s corporate group structure, I recommend this paper from law professors Curtis Milhaupt at Columbia Law and Li-Wen Lin from the University of British Columbia: We Are the (National) Champions). China’s corporate groups typically center around a single industry, in contrast to South Korea’s own national champions – the chaebol – which are diversified conglomerates bound in a complicated web around a founding entrepreneurial family. So, subsidiary firms in this vertically integrated corporate group structure often act as either the suppliers or final customers for their listed parent company.

This type of intragroup related party transaction (RPT) is enforced by political incentives rather than arms-length contracts which, of course, reduces the demand for HQFI.

Related party transactions are particularly common among state-owned firms in China. For example, the 2015 annual financial report of the listed company in Hesteel Group, a Chinese iron and steel manufacturing conglomerate, reports purchasing iron ore, coal, and other input materials from forty-five other firms within the corporate group. Additionally, Hesteel Group’s listed company sold finished product to sixty firms from within the group. Just how significant is that? The cash value of RPT between Hesteel and its customers and suppliers total 88% of operating revenue.

The above chart shows the proportion of non-bank non-petrol firms conducting related party transactions as well as the proportion of the cash value of RPT to total revenue by ownership class. State firms are more likely to conduct related party transactions than private sector firms, and when they do, the amounts involved are greater than for private sector firms. In sum, the corporate group structure used by SOEs opens the door for extensive use of related party transactions which in turn diminishes the demand for HQFI.

Fifth and finally, is graft. From Maotai-soaked banquets, to luxurious travel with mistresses, to Ferrari-crashing scions, it is an open secret – widely panned by China’s netizens – that the behavior of government officials isn’t always beyond reproach. To quote Jamil Anderlini at the Financial Times, “Executives at China’s enormous state energy monopolies command vast armies of employees and control budgets worth hundreds of billions of dollars. They are often accused of acting like rulers over mini-states within the state.” Reported in the same piece, since the start of 2014 the Chinese government has detained over 100 top executives on suspicion of corruption.

Obvious to the point of banality, it is in the best interests of those engaged in rent-seeking activities to suppress information uncovering their behavior. When, as in the case of some top SOE executives, firm resources are appropriated for personal benefit, the incentives for obfuscation far outweigh those for transparency and the production of high quality financial information suffers.

Although government ownership of listed firms impacts the incentive structure surrounding data quality via several important mechanisms, the government impacts data quality through more than just ownership. Heavy-handed intervention in China’s capital markets also plays a major role.

Institutional Arrangement II: Government Control of Capital Markets

Here are Carl Walter and Fraser Howie in Privatizing China on the true purpose of markets in China:

Securities markets in any other country in the world evolved in market economies in support of, and benefitting from, the capital-raising efforts of privately owned companies. In China, stocks and markets had to be somehow squeezed into an economy still based on state planning and the absence of private ownership. As a result, the stated objective of securities markets has been to promote greater operating efficiencies in SOEs still controlled absolutely by the state. It is, therefore, entirely unclear just what securities do, in fact, represent, other than the opportunity perhaps to profit from trading and to receive an occasional dividend.

In other words, the government uses markets as just another administrative tool for distributing resources. Government control of capital markets impacts data quality in several ways – firms are incentivized to manage earnings, they gain little from transparency, and institutional investors, restricted in their arbitrage activities, lack the incentives to produce and trade on firm-specific information.

First, the China Securities Regulation Committee (CSRC) relies on bright line accounting regulations to manage resource allocation on China’s public equity exchanges which incentivizes firms to get creative with their accounting practices.

One example of the strict regulations imposed by the CSRC is that firms posting consecutive years of losses are relegated to ‘special treatment’ status, meaning they are ineligible to issue additional equity and daily trading is confined to a price window of plus or minus five percent. If in the third year the firm still hasn’t returned to profitability, the CSRC will delist them. End of discussion. Another example is that any firm hoping to raise capital via a seasoned equity offering (SEO) must first clear a threshold of six percent return on equity to receive CSRC approval.

These rigid bright line regulations incentivize firms to engage in earnings management which dramatically impacts the quality of financial reporting in China.

Let me draw your attention to two things on this chart. First, to stave off delisting, Chinese firms massage earnings by booking ‘big-bath’ losses in one year and profits in the next. Second, to receive government approval for an SEO, firms manage earnings to boost themselves over the 6% ROE threshold. That is why we see so few firm year observations below 0% ROE as well as a dip below 6% ROE. We can observe both mechanisms directly with a closer inspection of ROE data.

The left-hand side of this chart uses a box and whisker plot to visualize the distribution of return on equity values from 2013 to 2015 of firms that were unprofitable in 2014. This slice of data, while not historically comprehensive, provides a stylized illustration of management’s accounting decisions – a clear pattern of modest profitability in the years bookending a single year of substantial loss; this implies that firms make use of the creative accounting of accruals and related party transactions to combine and report three years of losses in just one year.

The right-hand side of the above chart shows the median return on equity for firms in the two years preceding and following a seasoned equity offering. As expected, we see a substantial drop in profitability the first year after an SEO. One note, an additional incentive for managers considering an SEO is for firm valuations to be as high as possible – that is why there we see an additional cushion above the 6% cut-off line, this incentive structure and rational firm response has been observed in developed markets as well.

So, CSRC regulations clearly incentivize firms to manage earnings around a unique set of circumstances. This inflicts a peculiar damage to the quality of financial data coming out of China. (For those interested, here is a forensic accounting manual that teaches the detection of uniquely Asian financial irregularities.)

The second mechanism of government intervention in capital markets reduces the demand for HQFI by obviating the benefits that could accrue to firms with highly transparent disclosures. In developed markets, borrowers with a track record of timely and transparent financial disclosures are rewarded with a lower cost of capital. But, the information needed to make capital raising, financing, and listing decisions in China is often obtained via non-financial channels such as political networks.

A very recent example of this was the Postal Savings Bank of China IPO in Hong Kong. Three quarters of the shares for sale were subscribed by mainland SOEs like China Shipbuilding Industry Corporation and the State Grid Corporation – all behemoths of the ‘old economy.’ Given that the Postal Savings Bank was one of the only large Chinese banks to trade above book value at the time, this IPO valuation was clearly a political rather than market decision.

We’ve been looking mainly at China’s equity markets so far, but an example from the corporate bond market illustrates clearly how the information required for financing decisions is obtained through political networks. State agents populate the ranks of both buyers and sellers of debt in China’s corporate bond market. In fact, more than 90% of the current outstanding corporate bond balance was issued by either SOEs or local government financing vehicles. Additionally, state-run banks and ‘Special Clearing Members’ such as the Ministry of Finance, policy banks, and the central bank, combine to own nearly a third of the outstanding corporate bond balance.

In a market where both the sellers and buyers of corporate bonds in China are state agents, the allocation of credit is an essentially administrative exercise. So, there is little incentive for firms to be fully transparent in their financial disclosures which would grant them access to lower cost capital in a market-driven environment.

The third and final way that China’s government control of markets curtails demand for HQFI are the numerous market frictions that hinder institutional investors from engaging in arbitrage activities. Short sale constraints, circuit-breakers, and T+1 settlement all prevent informed investors from profiting off of superior information. This contributes to the commodification of Chinese equities and high stock price synchronicity.

So far, we have explored how government ownership and interference in China’s markets disincentives market participants from disclosing, producing, or demanding high quality financial information. However, the most problematic aspect of China’s institutional arrangements is the legal system.

Institutional Arrangement III: Weak Legal System

There’s no point in going to court in China. It will be a long, expensive, frustrating ordeal, and you will probably lose. For those used to the protections offered by the US legal system, China can be quite vexing. A chilling example, although not directly related to financial information quality, can be found in the results of the criminal judicial system.

In 2015, 99.92% of all criminal defendants in China were found guilty. For a sense of scale, you should know that the raw numbers were 1,039 acquittals out of 1.23 milliondefendants. While this appalling example raises the ire of international human rights groups, it also serves to illuminate how the fundamental nature of China’s legal system is problematic for financial data quality.

Namely, China is an authoritarian state – a country where the government accepts no external checks to power. This has a second-order, but serious, negative impact on financial data quality.

To quote Piotrowski again:

There is no independent judiciary in China and the courts are controlled by the local governments, which also directly or indirectly hold the stock ownership of the state firms. Therefore, the courts do not and cannot serve as the ultimate arbiter in legal disputes.

In this environment, firms and investors alike are unable to turn to the courts to enforce contracting agreements. Why? Because they will lose. Every time. So, contracting parties must enforce the use of contracts by depending on their social and political networks – that is, you do business with your classmate, brother-in-law, or, as in the case of the Hesteel Group, a subsidiary company. Firms do this because the enforcement of such related party contracts doesn’t depend on the court system.

Given these weak legal institutions, the rational response for Chinese firms to minimize agency conflicts is to consolidate both ownership and management. In the case of SOEs, as explored earlier, the state minimizes the shareholder-management agency conflict by consolidating ownership. In the case of the private sector, tightly-knit family businesses with overlap between management and ownership are the norm.

The left-hand side of the chart shows that Chinese firms have highly concentrated ownership. In fact, the median shareholding percentage for the largest shareholders in China is nearly three times higher than for New York Stock Exchange traded firms. The right-hand side of the chart illustrates the overlap between management and ownership in China’s private sector. The controlling shareholder for 79% of private sector firms also serves as the CEO or chairman of the board for that same company.

This highly concentrated ownership structure grants shareholders some measure of security in the absence of an independent judiciary. That is, it allows shareholders to monitor the firms through direct and informal channels – outside of accounting disclosures. Of course, the inevitable side effect of which is an attenuated demand for HQFI.

Prospects for Improvement: Partly Cloudy

China’s regulators have worked to improve the information environment for investors. For example, Chinese accounting practices have already achieved 95% convergence with IFRS. However, despite these surface level similarities, there are some major obstacles impeding true improvement in the information environment.

First, as mentioned before, China is an authoritarian state. That is, the government suffers no checks from external sources and completely controls the court system. So, it is unlikely that China’s legal institutions will ever evolve to a point that gives firms and investors the protections necessary to incentivize the production and disclosure of HQFI.

Second, the Party is unlikely to let go of ownership of listed firms. Remember, China’s stock markets exist as an administrative tool of the government and many of the political elite have become wildly wealthy from their involvement with SOEs. This will continue to hamstring the production and disclosure of HQFI.

In the medium term, we can hope for two developments that will improve China’s information environment. One, if the number of private sector firms listed on the stock exchanges continues to grow, then firms which are properly incentivized to produce and disclose HQFI will slowly begin to outnumber those inclined to opacity. Two, if the government scales back the heavy-handed intervention in the markets which inhibits smart money from trading off superior information, then institutional investors will be incentivized to produce firm-specific valuable information.

A Bright Spot: Chinese Analysts

In all this mess there is one bright spot – analysts; they exert a positive influence on the information environment of China’s capital markets.

That being said, China’s analysts aren’t without their flaws. Sell-side analysts in general, and Chinese analysts in particular, have a tremendous conflict of interests between their investment banking and brokerage businesses. Therefore, earnings forecasts typically skew overly optimistic to encourage clients to buy and to keep management happy. Additionally, of all Chinese listed firms that received an investment rating from local analysts, only seven were given a sell rating. Seven.

Despite their flaws, China’s ‘star’ analysts have been proven to actually do a good job of information production – that is, reducing stock price synchronicity for firms under coverage.

Finally, Chinese analysts can help you make money. A study from the Shanghai Advanced Institute of Finance found that a terribly naïve long-only strategy which buys stocks the day after receiving a “strong buy” rating and holds for 63 days will generate a positive annual alpha – after accounting for size, momentum, and value effects – of 11.60%. Additionally, the study finds that analyst recommendations are particularly powerful predictors of future returns for smaller firms and those with low analyst coverage. In other words, Chinese analysts produce useful information to help investors make money.

Conclusion: Where Do I Go from Here?

Chinese data is bad. But, if you haven’t fallen asleep yet, then you know that there is much more to the story. A very specific set of institutional arrangements in China incentivize rational market participants to behave in ways that circumscribe the quality of financial information. Understanding what those mechanisms are and how they work should give investors a more nuanced understanding of the information environment to avoid throwing out the baby out with the bathwater.

So, what can you do about it? Allow me to submit two helpful, albeit simplistic, conclusions following from an understanding of the institutional arrangements and incentive structures of China’s market participants. First, we can construct a checklist for assessing the data quality of Chinese listed firms – higher quality auditors, more diffuse ownership, and limited state involvement all signal HQFI. Second, we have developed an intuition for what data looks like when it is bad – related party transactions, earnings management around China’s unique set of incentives, and limited information creation from institutional investors all signal poor quality financial information.

My hope is that I have helped you begin to hone an intuition for evaluating with subtlety and rigor the quality of Chinese financial information, a skillset vital to all strands of China watchers.