Until recently, Chinese domestic stock markets, so-called A-share markets, were largely shunned by international investors. But as China recovered faster than most countries from the sudden economic contraction induced by the Covid-19 pandemic in the first quarter of 2020, and remained one of the few economies expected to grow this year, the tide is appearing to turn for these mainland stocks.
As we enter into 2020’s final stretch, A-shares are amongst the best performers year to date, with strong double-digit returns. Looking forward to 2021, this situation remains relatively favorable. Although a re-escalation of US-China tensions is possible, if increasingly less likely, the Chinese economy is expected to post significant growth, helped by strong fiscal and monetary support as well as ongoing structural reform efforts.
Yet this does not mean investors should rush headlong into buying. Although big steps have been taken over the past decade, A-share markets remain in their early development stages. Regulation is still catching up with common standards in developed markets, and important issues still exist regarding disclosure and transparency. This is particularly relevant for quantitative investors, as their models – usually designed to operate in developed markets – typically fail to capture such risks.
Fraud cases are by no means a marginal phenomenon in China, as Figure 1 illustrates. It shows a steady rise in the number and type of regulatory violations reported by the China Securities Regulatory Commission (CSRC) between 2001 and -2019. Since 2015, the number of enforcement actions has exceeded 400 per year. The most common types of violation are shown separately, being delayed disclosure, false recordation, and material omission.
Figure 1: Rising number of regulatory violations reported by the CSRC
Source: Robeco, China Stock Market and Accounting Research. We count the number of regulatory violations reported by the CSRC for each calendar year.
Admittedly, the number of listed firms has risen sharply over this period, especially since 2012, which may partly explain why fraud cases have been growing sharply. Also, the frequency of CSRC enforcement actions has also risen, relatively speaking.
An important point to remember is that fraud cases are not only restricted to small companies. Among constituents of the CSI 300 Index, the 300 largest A-share companies, there have been at least 20 fraud cases every year since 2014, with a high of 35 cases in 2015.
Opportunity and incentive
The two main conditions necessary for fraud to prosper are opportunity and incentive, both of which tend to be high in China. The country’s still rudimentary regulatory framework means that Chinese companies tend to have weaker internal controls and poorer governance than their Western counterparts. Moreover, sanctions for regulatory violations are often relatively minor, and the potential benefits from undetected fraud can be huge.
In May 2019, for instance, a Chinese pharmaceutical company was found to have falsified its financials. However, the firm’s chairman was not sentenced to a jail sentence and was only fined RMB 0.9 million (EUR 120,000), after having pocketed RMB 10 billion (EUR 1.3 billion) from the fraud. Meanwhile, the stock was allowed to resume trading only after a short suspension period, although its price subsequently declined by more than 80%.
As a comparison, back in 2009, Bernie Madoff was sentenced to no less than 150 years in prison in the US (the maximum allowed), having been found guilty of stealing an estimated USD 65 billion from approximately 4,800 clients, through the largest Ponzi scheme in world history, making it the country’s largest financial fraud case to date. Such harsh sentence reflects the much stricter approach to fraud sanctions applied in the US and more generally in Western countries.
As long as opportunities remain plenty and incentives high, the number of fraud cases is likely to remain substantial in Chinese A-share stock markets. It is therefore crucial for quant investors to have the detection tools necessary in order to avoid such risks. Unfortunately, existing available databases relating to fraud-risk screening lack the broad coverage and level of detail typically needed for this purpose.
A proprietary fraud-risk screen
To address this issue, Robeco has built a proprietary fraud-risk screen to cover the entire investment universe of approximately 1,500 China A-shares that we rank for our quantitative equity strategies. The aim is to first identify which companies have questionable business and accounting practices, before any potential fraud is exposed by regulators or the media. These practices might result in low integrity of the data we use as input in quantitative models.
Companies that score high on our fraud-risk screen are then scrutinized in more detail, using fundamental analysis to determine the likelihood of low-integrity data reporting or potential regulatory violations. In our analysis, we consider approximately 100 company-specific variables, including accounting metrics. Table 1 explains some of the main variables we look at as part of our screening process.
Table 1: Five key variables explained
This variable measures the incentive for insiders to manipulate the share price. Pledged shares serve as collateral for loans. Insiders with large equity stakes in a company who have pledged their shares for loans to fund their stakes typically do not have much additional cash. If the share price drops, additional cash collateral is usually required by the lender. Hence, there is a strong incentive for insiders to manipulate stock prices when these are under pressure.
Verification whether the reported strong cashflows are genuine or overstated on the balance sheet. An important difference between genuine and fictitious cash is that the latter cannot be paid out as dividends to shareholders. Thus, companies with lots of reported cash but no dividend payments can appear suspicious. Also, artificially generated cash generally appears in the balance sheet in one form or another as a ‘non-production’ asset. These ‘non-production’ assets are assets which are not involved in the production of goods and services, and typically include short-term investments and accounts receivable. Companies with an abnormal level of ‘non-production’ assets, or an unusually fast build-up of ‘non-production’ assets over a short period of time, should also be considered with caution.
This highlights companies with poor capital management practices. Companies that report to be in a net positive cash position and generate strong free cashflows would normally not need to issue significant amounts of debt, nor raise new equity frequently. Also, dividend payments or share buybacks can sometimes be funded by debt. While this is not uncommon, it sometimes reflects poor corporate governance or financial statements falsification.
This aims to identify companies which manipulate their reported profits through a frenzy of acquisitions. Companies that make material acquisitions have more opportunities to manipulate their earnings. These companies can artificially boost earnings in the periods following acquisitions through asset write-downs during the M&A process. Asset write-downs lead to losses during the M&A process itself, but can artificially inflate accounting profitability metrics going forward through lower depreciations. This earnings manipulation may be used to fulfil post-merger earnings targets that may be part of the management’s incentive scheme.
The goal is to identify whether a company presents favorable financial statements by understating the actual level of debt. Although window dressing does not always involve fraudulent practices, it is usually done to mislead investors. Managers may engage in window dressing by reducing short-term borrowings to disguise the true risk level of the company. In this example they will often sell short-term speculative assets and use this money to repay short-term debt or fund other operations just before the reporting period. After the reporting date, they may again purchase speculative investments financed by short-term debt. The goal is to hide certain risky activities from investors.
Thanks to this fraud-risk screening tool, we can identify the dozen companies showing dubious management and accounting practices, out the hundreds of large and liquid investible Chinese A-shares. Robeco’s fundamental analysts in Rotterdam, Shanghai and Hong Kong can then study these companies in more detail to further narrow down the list of firms that are most likely to suffer from fraud.
Short-listed individual cases are ultimately discussed within the Quantitative Equity Department. Depending on the outcome, overweight positions may be cancelled, or stocks may be entirely excluded in our quantitative equity portfolios. A potential reason to decide not to entirely exclude a riskier company from all our strategies is that excluding large index names may result in a large tracking error, which can be a problem for some investors.
In a stock market characterized by very attractive investment prospects but also relatively high chances of being affected by fraud, our proprietary screening tool helps us remove important tail risks that are typically not captured by standard quantitative investments models. This fits well with our key investment principle to avoid unrewarded risks at each stage of the investment process, to ensure the highest risk-adjusted returns for our clients over the longer term.