In this literature review, our purpose was to identify how far research has gone about risk and about the specific risk of the Chinese market in order to define the angle of our own study. In a first part, we focused solely on the notion of risk, with the help of books and research papers defining different hypothesizes about the functioning of the market and the measures used by investing firms to reduce their risk. We also tried to uncover different notions of risk, from probability to uncertainty. The second part focuses on the specificity of China, its political and economic structure and how it makes China’s market unique. After an analysis of the situation we focused on studies in which the researchers found anomalies specific to China’s stock market and comparisons with the US’s stock market. In a third part, we define how this helped us find an approach to work on and enhance the comprehension of risk on Chinese stock market.
- Risk: definition, understanding, measurement, uncertainty
First, we can work on the notion of risk and the definition of the concept of risk. There are two components in the risk: the probability of an event, and the potential outcome. We can say risk is an exposure to uncertainty. Academic work on the risk in stock markets and stock prices started in the 1950s (Markowitz, Lintner, Sharpe, etc…) and lead to the creation of portfolio theory based on Capital Asset Pricing Model (CAPM) and Arbitrage Pricing Theory (APT). In 1988, the Basel Capital Accord raised a new concept: the regulatory risk. It led to a formalization a risk measurement for banks. It also means that it the method applied is wrong, everyone is wrong.
Then, we should define market risk. It is the risk of losses induced by changes affecting a public market and not a specific asset. The main factors of market risk can be listed as:
- Change in equity prices
- Change in interests rates
- Change in foreign exchange rates
- Change in commodity prices
There are several instruments and variables used to measure and report market risk. Among then, the most important one over the past 15 years has been the value-at-risk (VaR), it has been implemented as the basis in risk regulation (Basel II) and measurement, even if events during the period proved the limits of our risk management, as it happened in 1998 with Long Term Capital Management, in 2007-2008 when banks recorded more than 30 days when losses exceeded VaR instead of less than 5, and in 2011 when risk models rated eurozone governments bonds as almost risk-free just before the European sovereign crisis. Trying to cope with this, financial institutions have increased the number of risk factors required to price an asset, or a trading book to several thousands, creating increasingly complex valuation models and integrating new stress-testing procedures to anticipate a large possibility of macroeconomic changes.
Cornelis Los, in his book Financial Market Risk: Measurement and Analysis, thinks the best way of measuring risk is as a residual, unexplained irregularity, which require to analyze fractality, or self-affinity, of speculative and cash market pricing, and propose various forms of measurement and visualization of long-term dependence, of market persistence or antipersistence, using wavelet Multiresolution Analysis (MRA). Two hypotheses are considered here:
- Efficient Market Hypothesis
- Fractal Market Hypothesis
Most textbooks in finance have blindly adopted the Efficient Market Hypothesis (EMH). However, the problem is still open. Several models have been proposed. For example, the Fractal Market Hypothesis (FMH), by Mandelbrot, which emphasizes the empirically observable and well-corroborated self-similarity of power laws of financial pricing contradicts the EMH based on martingale theory.
FMH: The magnitude of a risky market asset pricing process is both frequency (=asset class ω) dependent and horizon τi-dependent and shows global dependencies via its self-similarities in the frequency and time domains. Such a process can be homogeneous = mono-fractal (=exhibiting one form of self-similarity), or non-homogeneous = multi-fractal (=exhibiting many coexisting self-similarities).
A geometric object is fractal when it has a fractional dimension D. Euclidian geometric objects have discrete integer dimensions, such as D(point) = 0, D(straight line) = 1, D(plane) = 2, D(cube) = 3. Fractal geometry describes objects that are self-similar. When such objects are magnified, their parts are seen to bear an exact resemblance to the whole, the likeness continuing with the parts of the parts and so on to infinity, as in the figure 1 below (Julia set).
Fractals are scale-symmetric (=self-similar): the fractal process can be magnified, but keeps the same shape. In financial risk-theoretic terms, a marginal distribution of a fractal speculative pricing process maintains its shape when observations are taken at different time intervals, although the size of these shapes (=their amplitude) can change.
Fractals are translational-asymmetric: the fractal process does not keep the same shape when shifted. Fractal are devoid of translational symmetry, i.e., they do not exhibit the smoothness associated with Euclidian lines, planes and spheres. Instead a rough, jagged quality is maintained at every scale at which an object can be examined.
Why is the study of fractals so important? The classical financial models used for most of this century handles extreme situations with benign neglect: it regards large market shifts as too unlikely to matter or as impossible to take into account. It is true that portfolio theory may account for what occurs 95 percent of the time in the market. But the picture it presents does not reflect reality, if one agrees that major events are part of the remaining 5 percent. An inescapable analogy is that of a sailor at sea. If the weather is moderate 95 percent of the time, can the mariner afford to ignore the possibility of a typhoon?
The risk-reducing formulas behind portfolio theory rely on a number of demanding and ultimately unfounded premises. First, they suggest that price changes are statistically independent of one another: for example, that today’s price has no influence on the changes between the current price and tomorrow’s. As a result, predictions of future market movements become impossible. The second presumption is that all price changes are distributed in a pattern that conforms to the standard bell curve. The width of the bell shape (as measured by its sigma, or standard deviation) depicts how far price changes diverge from the mean; events at the extremes are considered extremely rare. Typhoons are, in effect, defined out of existence.
The study of fractals gives a new understanding of the changes in stock price, which can be used to stress-test portfolios and maybe lead to a new wave of risk models.
In his book The Handbook of Market Risk, Christian Szylar produced a great presentation of the field, helping reduce complexity and improve understanding of market risk. More than that, it shows how all the work and perception of risk which was common knowledge before the 2007 crisis had to change.
The particularity of this crisis was its amplitude and severity, second only to the crisis of 1929, and the progress that had been made in the past decade in the study of risk. Suddenly all the models and assumptions used in the investing and risk management environment were wrong:
- Capital requirement for financial institution was not enough to protect against bankruptcy.
- Stress testing was faulty and not run properly.
- Links between some OTC products were misunderstood.
- Liquidity risk was not properly monitored.
- The value-at-risk (VaR) method had created a false sense of security for all market participants.
- Correlations had not been managed properly.
- Over-reliance on some mathematical models.
- Risk governance was not working properly in many financial institutions.
We could quote Socrates saying, “I know that I know nothing”, as a lesson from this period. Investors thought we knew a great deal about market risk and its implications for investing firms. They realized afterwards that in fact, we did not know that much. And they discovered that the parameters were evolving. For example, diversification, a great tool to diminish market risk, has become extremely difficult. The markets have become obsessed with the idea of risk on-risk off, polarizing the market in such a way that correlation between different assets has become higher than ever. Usually correlations rise during crisis and fall back when the crisis has passed. However, in today’s world we are experiencing high correlations with low levels of volatility in a growing economy, which could suggest a structural change in the market.
A brief history of market events tells us something critical about our understanding of risk: every time we find a new way to measure risk and manage it, if it seems to work, we follow the model blindly until a market event shows how wrong we were. As long as things go well, why should we change what we are doing? An equivalent reasoning would be playing Russian roulette and assuming nothing happens when you pull the trigger it is considered safe to do it again. But it is exactly the way the business is working. It looks like progress is made only because market events tell us we should look for something that works better. It is unfortunate but normal that risk management has evolved alongside economic catastrophes. We have tried to improve the system with the new knowledge brought by such events. For example, in 1933, to avoid a new occurrence of the Great Depression, the US Congress passed the Glass-Steagall Act to prohibit common ownership of banks, investments banks and insurance companies.
We should also remember that past events that we anticipate now as a risk factor were at that time completely unexpected, unpredictable. Frank Knight taught us in Risk, Uncertainty and Profit to be careful about the difference between risk and uncertainty. We can measure risk, we cannot measure uncertainty, its influence is a surprise, which means we should not be over-reliant on past frequencies in order to predict the future. Following Frank Knight’s writings, Kenneth Arrow won the Nobel Prize in Economic Sciences in 1972 by imagining a “perfect world in which every uncertainty is ‘insurable’, a world in which the law of large numbers works without fail”. He then pointed out that our knowledge is always incomplete and we are best prepared for risk by accepting its potential as both a stimulant and a penalty, not by trying to eliminate it.
In his paper The Psychology of Risk, Amos Tversky explains us a key point about the issue of risk modeling. An assumption that we did not think about and might be wrong, the assumption that the individual is rational. It appears that people are hindered by many factors, some of which are emotional and some of which are cognitive. Herbert Simon has called it “bounded rationality”. The psychological analyses of judgement and choice have uncovered several phenomena that contrast sharply with the basic tenets of the standard theory. In particular, they have shown that people often exhibit risk seeking, that they tend to segregate the outcomes of different decisions, and that their expectations are often biased in predictable directions.
Most models of investment behavior are based on the assumption that investors are risk averse. Risk aversion does not always hold, however: When losses are involved, people often exhibit risk-seeking behavior. This behavior can be explained in terms of an aversion to losses.
The difference between risk aversion and loss aversion can be illustrated with an example. Consider the case where an individual must choose between two opportunities to gain: a sure $85,000 or an 85 percent chance of receiving $100,000 and a 15 percent chance of receiving nothing. Faced with this choice, the majority of people prefer the sure option, even though the expected return is the same in both cases. Apparently, they feel that $85,000 for sure is better than a chance of getting a little bit more and a non-trivial chance of getting nothing. This choice reflects risk aversion: a sure outcome is preferred over a gamble with the same expected value. Now consider the opposite problem: One has to choose between a sure loss of$85,000 oran85 percent chance of losing $100,000, and a 15 percent chance of losing nothing. In this situation, most people prefer the gamble over the sure loss. Evidently, they feel that a sure loss of $85,000 is worse than the risk of losing a little more with the gamble, combined with a chance that they will not lose at all. So in situations where the probability of loss is quite large, people exhibit risk-seeking rather than risk-averse behavior: They find the gamble more attractive. This is an example of loss aversion. This kind of phenomenon demonstrates that human judgement and choice are subject to serious cognitive limitations. People who plan to use investor psychology in their investment approach should not ignore these limitations.
In Financial Market Risk: Measurement and Analysis, Cornelis Los tries to answer two questions:
- Is all financial market risk dangerous or can we distinguish between “safe” financial market risk and “dangerous” financial market risk?
- How can we manage financial market risk to our advantage?
A part of the book, perhaps the most interesting for us, is about the phenomena of chaotic risk and financial turbulence. It defines financial chaos and demonstrates how such chaos can develop in financial markets. Financial turbulence is an efficiency enhancing phenomenon that only occurs in the antipersistent, most liquid anchor currency markets. Financial turbulence should therefore be distinguished from financial catastrophe or crisis.
The last part of the book discusses Extreme Value Theory and the consequences on the Value-At-Risk approach to portfolio and bank management in order to reduce financial risk. But sometimes financial market risk cannot be reduced because of its peculiar empirical characteristics of long-term time-dependence and non-stationarity, a phenomenon already studied in the 1960s by Fama and Samuelson.
According to Los, the financial markets suffer from three major deficiencies:
- Risk is insufficiently measured by the conventional second-order moments (variances and standard deviations). This relies on the facile, but erroneous, assumption of normality (or Gaussianness) of the price distributions produced by the market processes of shifting demand and supply curves.
- Risk is assumed to be stable and all distribution moments are assumed to be invariant, i.e., the distributions are assumed to be stationary.
- Pricing observations are assumed to exhibit only serial dependencies, which can be simply removed by appropriate transformations, like the well-known Random Walk, Markov and ARIMA, or (G)ARCH models.
We have learnt that uncertainty of the investments is a much wider concept than just the volatility of the prices as measured by second-order moments. Higher order moments, like skewness and kurtosis, play an underestimated, but very important role. For example, the return distributions are positively skewed because of confirmation biases. Only surviving businesses are shown in the distribution.
For physicists, uncertainty is fundamental. The world could not exist without uncertainty. “One the fundamental consequences of uncertainty is the very size of atoms, which, without it, would collapse to an infinitesimal point.” (Schroeder) In modern risk theory, we distinguish three different but closely related concepts: randomness, chaos, and probability.
Randomness refers to irregularity. Chaos is a special form of irregularity. It means that at a certain time something that was certain and unique, suddenly become nonunique, although it remains very well determined. Complete chaos is the coexistence of an infinite number of unstable deterministic equilibrium orbits through which the dynamic system cycles. It is a form of inexactness. Probability refers to the complete set of relative frequencies. It is a very well defined and constraint form of randomness. The basic problem of using a probability measure to describe a degree of uncertainty is that one has to know the complete universe of states that may occur to compute the relative frequency or probability of a particular event. But the definition of the concept of uncertainty already indicates that we are doubtful or ignorant of what may happen. Thus, probability only plays a role in games that have completely predefined rules. Most real-life situations are not like well-defined games. The probability distribution of a particular event is itself a phenomenon that science tries to discover and identify. It is pseudoscience to presume and predefine such probability distributions before the finite empirical data sets have been analyzed.
- China’s market: structure, uniqueness, anomalies
China’s stock market structure:
To have a better knowledge of the specific subject we are studying, i.e. the risk in Chinese capital markets, it is important in our opinion to highlight the uniqueness of China in the world of financial markets. By aggregating its three largest stock exchanges, which are all among the 10 largest stock exchanges in the world, we realize Chinese capital market is second only to the United States. Here are the three largest components of China’s capital market, sorted from the largest to the smallest:
Shanghai Stock Exchange (SSE):
The largest stock exchange in China is the Shanghai Stock Exchange (SSE). It is the 5th largest stock exchange in the world, and the second largest in Asia, behind Tokyo Stock Exchange. The Shanghai Stock Exchange as we know it today was created in 1990. Its market capitalization accounts for almost $3.9 trillion, and 1041 companies are listed.
Stock Exchange of Hong Kong (SEHK):
The Stock Exchange of Hong Kong Limited (SEHK) is the 6th largest stock exchange in the world and 3rd in Asia, right behind the Shanghai Stock Exchange, with a market capitalization of $3.3 trillion. The SEHK has been created in 1914.
Shenzhen Stock Exchange (SZSE):
The second largest stock exchange in mainland China, third in the People’s Republic of China, the Shenzhen Stock Exchange (SZSE), was established on December 1st, 1990. It is the 5th largest stock exchange in Asia, the 9th in the world with a market capitalization of $2.2 trillion for 1420 companies. The purpose of the SZSE is mainly to list subsidiaries of state-owned companies.
There are not only three stock exchanges. There are also 2 regulators:
The securities and Futures Commission (SFC) of Hong Kong is the body charged with regulating the securities and futures markets in Hong Kong. The SFC was created in 1989 in response to the krach of October 1987. It is considered to be a branch of the government but it is run independently.
The China Securities Regulatory Commission (CSRC) is the main regulator of the securities industry in China, with Shanghai and Shenzhen stock exchanges under its jurisdiction. It was formed in 1992 as an institution of the State Council of the People’s Republic of China, and fully became the regulator of China’s stock exchanges after China’s Securities Law (passed in December 1998). Its headquarters are in Beijing.
This is not the only difference between Hong Kong and China’s other stock exchanges, Hong Kong Stock Exchange has been operating since 1914 when the other two started in late 1990. Trade in Hong Kong is made in Hong Kong dollars (HKD) and not in Chinese yuan (RMB). This unique situation finds its source in history. Before 1997 Hong Kong was a British colony, evolving alongside the Western world when mainland China was experiencing socialism. As a result, both places evolved apart from each other, with a different political system and a different currency. Hong Kong was handed back to China by the United Kingdom in 1997, but the difference being too important to merge Hong Kong system with the rest of the country, it was given the status of Special Administrative Region (SAR), leading to the saying “One country, two systems”.
Further than that, people can have questions about the functioning of a stock market, the most capitalist institution, with the doctrine of market socialism in a country run by a communist party, as in China. The book China’s Stock Market: A Guide to its Progress, Players and Prospects helps us to get a better understanding of its history and structure, and to extract the key points of its functioning, from the pragmatism of its creation to the share structure (A-shares / B-shares / H-shares), along with the political logic that has shaped the market’s development. First, the equity market is segmented by different types of shares:
- A-shares: common stocks denominated in Chinese Yuan and listed on the Shanghai or
Shenzhen stock exchanges.
- B-shares: special purpose shares denominated in foreign currencies but listed on the
domestic stock exchange.
- H-shares: offshore listed shares denominated in Hong Kong Dollars and traded on the
Hong Kong Stock Exchange.
- L-chips, N-chips and S-chips: shares of companies with significant operations in China,
but incorporated respectively in London, New York and Singapore.
- American Depository Receipts (ADRs): an ADR is a negotiable certificate issued by a
U.S. bank representing a specified number of shares in a foreign stock traded on an
Second, unlike most of the Western markets, China’s stock market is owned mostly by retail investors rather than institutional ones (around 80% in 2015).
Anyway, China is becoming a developed market, and it is developing fast. Many reforms were implemented during the past decade, and, in 2014, in order to make access easier to Shanghai Stock Exchange, the Shanghai Hong Kong Stock Connect was launched. All Hong Kong and overseas investors are allowed to trade eligible shares listed in Shanghai, but only Mainland institutional investors or individual investors who satisfy the eligibility criteria (i.e. Individual investors who hold an aggregate balance of not less than RMB 500,000 in their securities and cash accounts) will be accepted to trade SEHK Securities through Shanghai-Hong Kong Stock Connect. From Shanghai Stock Exchange, only A-shares are included in Shanghai-Hong Kong Stock Connect. B-Shares, ETFs, bonds and other securities are not included.
In a research paper from 2015, The Real Value of China’s Stock Market, Carpenter, Lu, and Whitelaw present evidences that overturn two widely held perceptions about China’s stock market and suggest that China would do well to open this market to international investors. First, they show that China’s stock market no longer deserves its reputation as a casino. On the contrary, over the last decade, the informativeness of stock prices about future corporate earnings has increased steadily, reaching levels that compare favorably with those in the US. Moreover, although China’s financial market is largely inaccessible to foreign investors, the cross-sectional pattern of its stock returns is strikingly similar to that in global equity markets. Like global investors, Chinese investors pay up for large stocks, growth stocks, and long shots, and discount for illiquidity and market risk. In addition, the trend of stock price informativeness in China over the last two decades is significantly positively correlated with the efficiency of its corporate investment. Stock prices in China have become strongly linked with firm fundamentals and appear to play an important role in aggregating diffuse information and generating useful signals for managers.
Second, although the buy-and-hold return earned by undiversified domestic investors in China’s stock market is depressed by the market’s extremely high volatility, the market offers very attractive returns to diversified international investors who can access them. Unlike stock returns in integrated financial markets, stock returns in China exhibit very low correlation with those in other large economies. At the same time, the average monthly excess return on China’s stock market is twice that in the US. As a result, China offers high alpha with respect to global risk factors to international investors who can access it.
However, this high alpha to potential international investors amounts to an inflated cost of equity capital for China’s firms, constraining the investment of its smaller and more profitable enterprises. In addition, the high volatility of China’s stock market represents high systematic risk to Chinese investors who cannot diversify it. Both of these problems would be mitigated by opening financial capital flows between China and the global investment community. Further reforms that liberalize capital flows and improve stock price informativeness will also be important to increase China’s investment efficiency and fuel its continued economic growth.
China’s financial system is dominated by its state-controlled banking sector and expanding shadow banking sector. The post-crisis expansion of China’s shadow banking sector is partly by design. This sector has been crucial to implementing China’s massive post-crisis economic stimulus by quickly channeling large amounts of capital to new real estate and infrastructure projects in order to avert a recession, fuel real investment, and sustain economic growth. However, the implicit guarantee of China’s shadow banking sector may be undermining the development of an equally important financing channel–China’s stock market
In the paper, they analyze the informativeness of China’s Stock Market over the period 1996-2012. Informativeness of the market is defined as the cross-sectional variation in future earnings predicted by equity market value. Their results suggest that the informativeness of prices has steadily improved. Then they examine the efficiency of corporate investment in China over the same period. They define the efficiency of investment as the unexpected change in equity value associated with a unit of unexpected investment, measured in a cross-sectional regression. They find that the trend of investment efficiency follows that of price informativeness over the sample period with a high correlation. It emphasizes the real economic value of China’s stock market. Having established this link, they turn to the study of stock returns in China. In contrast to the high correlations in returns across open markets, returns in China’s stock market exhibit a low correlation with those in stock markets in other large economies. Yet, despite China’s segmentation from other markets, the cross-sectional pattern of its stock returns is remarkably similar to that in the US and in other global equity markets. China’s stock market seems to be as efficient as those of other large economies. Finally, they analyze the overall performance of China’s stock market and uncover a number of striking new results. Despite the SSE reputation of poor performance during the post-crisis period, their analysis of the broader market suggests that China’s stock market has in fact done very well over the full sample period 1996-2012, and offers attractive returns and opportunities for diversification to international equity investors who can access it. They show that China’s stock market has had a high average monthly USD return, and a high alpha with respect to the US and global market, size, value, and momentum factors. However, this high alpha translates to a high cost of capital for China’s firms.
Important questions for future research include the role of China’s economic reform risk in global financial markets; asset pricing and corporate listing choice across international exchanges; intermediary asset pricing and capital allocation in China; and interest rate liberalization, government bond pricing, and macroeconomic risks in China.
In other studies, such as Calendar effects of the Chinese stock markets, Wang, Ojiako, and Wang demonstrated recurrent anomalies they named calendar effects:the turn of-the year effect, the month-of-the-year effect and the day-of-the week effect. The study employs observations drawn from the Shanghai and Shenzhen composite stock indices between January 2000 and December 2010. In the Shanghai stock market, there is significantly positive Monday effect and significantly negative Thursday effect. On the other hand, they find a weak Monday effect and Thursday effect in the Shenzhen stock market. For monthly effect, summary statistics of monthly equity returns for both stock markets show the findings of the February effect in China are similar to the January effect in countries where year-end is in December.
In a GARCH model, Zhang et al. (2008) tested six market indices of the Chinese stock market from 1992 to 2003, and find abnormal returns occurring in March, when China is in the political high season. To sum up, while there has been progress in the development and efficiency of Chinese stock markets, there are still observable calendar effects as evidence of market inefficiency.
Other anomalies stand out as well, which are not caused by superstition or government intervention this time. The latest one, pointed out by Fox Hu, Steven Yang, and Amanda Wang on Bloomberg, is definitely able to grab one’s attention:
The Shanghai Composite Index, notorious for its volatility, has gone 85 trading days without a loss of more than 1 percent on a closing basis, the longest since the market’s creation in 1991 (figure 2). On 13 days during the streak, the index recovered from intraday declines exceeding 1 percent to close above that threshold, sometimes erasing almost 1 percent in the final 90 minutes of trading. For some investors, it is a sign that state-directed funds are here to ensure a stabilization of the stock market. This resilience could also be explained by global factors. Equity volatility has dropped around the world this year to reach historic lows.
In another study, Anomalies in Chinese A-Shares, by Hsu, Viswanathan, Wang, and Wool, from 2017, they consider that much of prior research on anomalies in China’s stock market are due to the fact that it requires more than constructing simple long-short portfolios and measuring historical average returns and associated t-stats. First, because China’s domestic equity market offers a very short sample of returns and accounting data. Those data are subject to evolving reporting standards, numerous changes in regulations, and government intervention. Then, the segmentation of China’s market into multiple share classes with varying levels of liquidity complicates the matter. Furthermore, short sales only became possible in 2010, and trading suspensions are a common occurrence, which negatively impacts price formation and market liquidity. Finally, as retail investors account for 85% of all trades in the Chinese equity market, the potentially biased behaviors of individual investors are likely to have a more important impact on Chinese stock returns than in developed markets.
As A-shares have become accessible to foreign investors in 2002, following the Qualified Foreign Institutional Investors (QFII) program, the study focuses on A-shares. As a prelude, they defined what characteristics or factors would be tested based on their return predictability and the attention they received in the US equities anomalies literature (figure 3). In a first part, they sort A-shares and US stocks into decile portfolios, then forming long-short portfolios by purchasing stocks in the top decile on each characteristic and shorting those in the bottom decile. To ensure the stocks are sufficiently liquid they work with the top 80% of stocks in each country based on market capitalization. The rebalancing takes place annually or monthly, depending on the signal.
|Size||SIZEt||Natural log of firm size||Natural log of the firm’s total A-share market value of equity from CSMAR, as of the end of
April in year t. CSMAR also provides firm market value calculated according to the free float,
but we believe that calculating size based on total shares outstanding provides a clearer measure
of the firm’s market value, particularly after the Split-Share Structure Reform of 2005.
|Valuation||B/Pt||Book-to-price ratio||Book value of equity for year t-1, divided by its total market cap. as of the end of April in year t.|
|E/Pt||Earnings-to-price ratio||Earnings for year t-1, divided by its total market cap. as of the end of April in year t.|
|S/Pt||Sales-to-price ratio||Sales for year t-1, divided by its total market cap. as of the end of April in year t.|
|D/Pt||Dividend-to-price ratio||Dividends for year t-1, divided by its total market cap. as of the end of April in year t. We drop
firm-year observations for stocks not paying dividends in year t-1.
|Profitability||GPt||Gross profitability||(Sales – COGS) for fiscal year ending at t-1, divided by total assets at the end of year t-1.|
|OPt||Operating profitability||(Sales – COGS – SG&A – Interest expense) for fiscal year ending at t-1, divided by book value
of equity at the end of year t-1.
|Investments||ΔASSETt||Change in total assets||Percentage change in total assets from the beginning to the end of the fiscal year ending at t-1.|
|ΔBOOKt||Change in book equity||Percentage change in book equity from the beginning to the end of the fiscal year ending at t-1.|
|ACCt||Accruals||Following the definition of total accruals in Hribar and Collins (2002), earnings less operating
cash flows, for fiscal year ending at t-1, scaled by total assets at the beginning of year t-1.
|NOAt||Net operating assets||Operating assets less operating liabilities for fiscal year ending at t-1, scaled by total assets at
beginning of year t-1
|Risk||VOLt||Total risk||The standard deviation of daily stock returns over the previous year.|
|BETAt||Systematic risk||The slope coefficient from a regression of daily excess stock returns over the previous year
against corresponding daily returns for an internally calculated capitalization-weighted market
benchmark consisting of all available A-shares.
|I-VOLt||Idiosyncratic risk||The variance of residuals from a regression of daily excess stock returns over the previous year
against corresponding daily returns for an internally calculated capitalization-weighted market
benchmark consisting of all available A-shares.
|Returns-based||MOMt||Momentum||The stock’s cumulative return from month t-12 to month t-2.|
|REV-STt||Short-term reversal||The stock’s return over the previous month.|
|REV-LTt||Long-term reversal||The stock’s cumulative return from month t-60 to month t-13.|
First, while the full-sample U.S. results largely line up with those in the past literature, the performance of A-shares factor strategies is something of a mixed bag. Based on the full A-shares sample, beginning in 1995, they observe that some signals performed quite well (size, value, accruals, and reversals), others posted weak, albeit positive returns (profitability, NOA, and low volatility), and two popular factors seem not to have worked at all (asset growth and momentum).
Second, when evaluating A-shares over a more recent sample—using data beginning in 2008, which should be more representative of present market conditions after major financial and accounting reforms, they observe a number of interesting changes. While some signals produce roughly the same effect as before (gross profitability, NOA, and low beta), most effects are stronger (size, asset growth, momentum, and short-term reversal) or weaker (book-to-price, sales-to-price, accruals, low-volatility, and long-term reversal), and several anomalies actually show opposite performance when going from one sample to another (earnings-to-price, dividend yield, and operating profitability). These differences highlight the effect developments in things like accounting standards and the regulatory environment might have on factor strategy performance.
Finally, it is unsurprising that over the shorter 2008-2016 sample the statistical significance of most results is substantially diminished; now only size, change in book value, and short-term reversal show statistical significance in A-shares. Similarly, while most of the well-known anomalies exhibit strong returns over the full sample in the U.S., covering 1965-2016, when they evaluate those same U.S. factor returns over a simulated ten-year sample, using data taken from each ten-year sub-period over the last five decades, the results are also less robust. This simple comparison offers an argument against a rush to conclude that a lack of significance necessarily implies the absence of an effect.
The low-volatility effect is an interesting case. Based on estimates from the recent sample, we would need well upward of 200 years of data to achieve a decent p-value. It seems difficult to make the case that the lack of significance for the low-volatility effect in China is based solely on the length of the recent sample. For other factors—accruals and NOA, valuation-based signals since 1995, or price-based signals in recent years—the required sample length is reasonably low, suggesting that with more data they might expect some of these factors to present with higher statistical confidence.
Then the study focuses on how factor investing in China is different. First, they focus on the impact of regulation and reforms. As we know, China’s financial landscape has evolved tremendously during the past two decades. Major accounting reforms were implemented in 2006-2007 that brought its accounting standards toward compliance with IFRS. To evaluate the significance of accounting reforms for accruals-based factor investing in China, they provide factor returns for the classic accruals strategy, the variation on accruals tailored to Chinese stocks through the exclusion of loss firms, and the NOA strategy. They observe that for both of the accruals factors, performance does appear to abate in the period after reforms. On the other hand, they note point estimates for the accruals strategies are still positive in the post-2007 sample, while the NOA factor—which might be seen as a more comprehensive measure of earnings sustainability—actually delivers more significant abnormal returns after accounting changes. The analysis indicates the potential sensitivity of factor returns to China’s continuing financial development, while supporting a diversified approach to factor investing whose performance is robust to a changing environment.
Then it focuses on a second important parameter specific to China: State ownership. Even after decades of privatization, over 50% of A-shares market cap was classified as deriving from SOEs as of the end of 2016. Two basic differences between SOEs and non-SOEs have clear potential to affect the performance of our factor strategies. First, the shares of state-owned firms have traded at perennially lower valuation multiples than those of private listed firms in China. As such, sorting on value-oriented signals has the effect of tilting one’s portfolio toward SOEs. Second, a sizeable proportion of state-owned shares were historically non-tradable, and the state has little reason to actively trade the unrestricted portion of its shares. These static holdings result in the shares of SOEs exhibiting relatively low liquidity and, consequently, persistently low levels of volatility. This suggests one high-level approach to addressing differences between SOEs and non-SOEs in terms of the factors tested in Section 2 is to construct SOE-neutral A-shares factor portfolios.
We can sum up this study as follow: it suggests that a number of factor investing strategies function as well in A-shares as in the US, but several traditional factors show surprising results when applied to Chinese stocks. Explanations such as accounting and financial reforms and state-ownership have been proposed as they create the landscape that makes China’s stock market unique.
- Working Hypothesis
From our analysis of the literature, we understand that most of the literature on market risk is firm-oriented, it is made to try to measure and manage the risk of a investing firm, not to understand the risk of the overall market. This risk is studied more in papers relating to non-conventional theories, like the Fractal Market Hypothesis and the work of Mandelbrot, or studies about psychology from Tversky and Kahneman. On the other hand, studies about China, given the youth of the stock market, have been focused on the understanding of the functioning of the whole market and its differences with the US stock market. We can also notice that Hong Kong stock exchange is not included in the studies as, even with the return of Hong Kong to China, the Mainland and the Special Administrative Region remain completely different economically.
From our readings, we noticed some interesting findings which could lead to other studies. One paper in particular has caught our attention: Anomalies in Chinese A-Shares, by Hsu, Viswanathan, Wang, and Wool. It is a recent paper, from 2017, and their findings lay the ground to new ones. They discover differences in the results of factor investing in China compared to the US, and they suggested explanations but these have not been proved yet. They highlight that factor investing in China is different for two reasons. The first one is about the impact of regulation and reforms. The second one is State-ownership. A large portion of China’s stock market is made of state-owned companies. We already know from their research that these shares trade at lower valuation multiples and have historically been non-tradable. We would like to explore the differences in factor investing strategies given different constraints on the weight of state-owned companies in a portfolio. We believe we could collect the data from the China Stock Market and Accounting Research (CSMAR) database. Then we would have to define the in our sample the state-owned companies and he private ones, and to determine the period of the analysis. The longer would be the better, but as many state-owned companies were made of non-tradable shares at some point we would probably have to work on the data to define the right range.
Alford, Andrew W., and Alison W. Lau. “A Foreign Investor’s Guide to Accessing the Chinese Equity Market.” The Journal of Portfolio Management 41, no. 5 (2014): 31-40.
Bernstein, Peter, 1996. Against the Gods: The Remarkable Story of Risk. John Wiley & Sons.
Bouchaud, Jean-Philippe, and Marc Potters, 2000. Theory of Financial Risks: From Statistical Physics to Risk Management, Cambridge University Press.
Carpenter, Jennifer, Lu Fangzhou, and Whitelaw Robert, 2015. “The Real Value of China’s Stock Market.” The National Bureau of Economic Research.
Chorafas, Dimitris N., 1994. Chaos Theory in the Financial Markets. McGraw-Hill Education.
Green, Stephen, and The Economist, 2003. China’s Stock Market: A Guide to its Progress, Players and Prospects. Profile Books
Gungwu, Wang, and Wong John, 2007. Interpreting China’s Development. World Scientific Publishing Company.
Hilliard, Jitka, and Haoran Zhang. “Size and price-to-book effects: Evidence from the Chinese stock markets.” Pacific-Basin Finance Journal 32 (2015): 40-55
Hsu, Jason, Viswanathan Vivek, Wang Michael and Wool Philip, 2017. “Anomalies in Chinese A-Shares.” Social Science Research Network.
Hu, Fox, Yang Steven, and Wang Amanda. “China’s Stocks Refuse to Drop More Than 1%.” Bloomberg. 19 Apr. 2017.
Huang, Yujia, Jiawen Yang, and Yongji Zhang. “Value premium in the Chinese stock market: free lunch or paid lunch?.” Applied Financial Economics 23, no. 4 (2013): 315-324.
Hull, J. C., and A. White, 1998. “Value at Risk When Daily Changes in Market Variables are Not Normally Distributed,” Journal of Derivatives.
Knight, Frank, 1921. Risk, Uncertainty, and Profit. Hart, Schaffner & Marx.
Li, Bin, Judy Qiu, and Yanhui Wu. “Momentum and seasonality in Chinese stock markets.” Journal of Money, Investment and Banking 17 (2010): 24-36.
Los, Cornelis, 2003. Financial Market Risk, measurement and analysis. Taylor & Francis.
Luo, Jin, Gan Christopher, Hu Baiding, and Kao Tzu-Hui, 2007 “An empirical analysis of Chinese stock price anomalies and volatility.” Lincoln University.
Mandelbrot, Benoit B., 1997. Fractals and Scaling in Finance: Discontinuity, Concentration, Risk. Springer Verlag, New York.
Mandelbrot, Benoit B. “How Fractals Can Explain What’s Wrong with Wall Street”, Scientific American. 15 Sept. 2008.
Naughton, Tony, Cameron Truong, and Madhu Veeraraghavan. “Momentum strategies and stock returns: Chinese evidence.” Pacific-Basin Finance Journal 16, no. 4 (2008): 476-492
Patel, N. R., Radadia, N. and Dhawan, J., 2012. “An Empirical Study on Weak-Form of
Market Efficiency of Selected Asian Stock Markets.” Journal of Applied Finance and
Banking, 2(2), 1792-6599.
Szylar, Christian, 2013. Handbook of Market Risk. John Wiley & Sons.
Tversky, Amos, 1989. “The Psychology of Risk”. Quantifying the Market Risk Premium Phenomenon for Investment Decision Making conference, 26-27 Sept. 1989.
Wang, J., Ojiako, U., and Wang, L., 2012. “Calendar effects of the Chinese stock markets.”
International Journal of Business and Emerging Markets, 5(1), 67-82.
Zhang, Lin, 2014. China’s Venture Capital Market: Current legal problems and prospective reforms. Elsevier Science.
Zhang, Z., Sun, W. & Wang, H., 2008. “A New Perspective on Financial Anomalies in Emerging Markets: The Case of China.” Journal of Applied Financial Economics, 18, 1681-1695