ESG Trends

Alternative Data are increasingly used to Assess ESG Performance and Risks of Chinese Companies

Tracey Xiang 2019-11-21

The rapidly growing demand for ESG information of Chinese companies and data analytics services is being met by a rising wave of alternative data services assisted by artificial intelligence technologies.

The usage of and demand for non-traditional corporate information about Chinese companies, particularly environmental, social, and governance (ESG) data for investors to screen investments and manage ESG-related risks, is rising thanks to the combination of technology availability and the undergoing structural changes in China’s economy.

 

For many investors one of the major frustrations about investing in Chinese companies is insufficient or unreliable corporate information, either from corporate disclosures or other domestically located third-party sources. Lack of transparency and disclosure aside, the financial reporting and disclosure practices of Chinese companies are very different from those in the advanced markets like the U.S., given different institutional and legal environments, and social norms. For instance, Chinese companies always tend to be slower to disclose bad news than positive information.

 

Compared with the long-standing concerns on corporate governance and social impacts, environmental issues are a more recent concern for investors thanks to the changes in regulation, economic growth prospects and consumer attitudes and behaviors that took place in recent years. To combat environmental pollution and promote potential new economic drivers based on eco-friendly and energy-efficient technologies, China has come up with a series of regulations to crack down polluters and policies to promote financing towards sustainable and environmental friendly companies and projects. But the overall levels of regulatory compliance have been low, with a regulatory inspection conducted in 2017 revealing that 70% of companies they examined violated the pollution rules. So when it comes to environmental impacts, many Chinese companies are facing high regulatory and economic risks.

 

It’s hard to get reliable information on ESG issues from traditional sources in China. ESG disclosures haven’t been mandated yet and there aren’t national standards on ESG disclosures. It was only a couple of years ago Chinese regulatory authorities began to require public companies to disclose information on the environmental issues. As of June 2019, less than 26% of companies that were listed on local stock exchanges issued CSR (corporate social responsibility) reports for the previous year. And self-reported information by Chinese companies and many local news sources have long been scrutinized for reliability thanks to relatively high levels of corporate frauds and dark arts of public relations. (link in Chinese)

 

At the same time, information about companies or specific industries is exploding on the web and other digital channels. Fraudulent activities and other malpractices are increasingly being exposed and at a faster pace on social media and other online platforms. Companies and executives use increasingly more online channels, particularly social media platforms, to release information or post content to engage different audiences. Information as obscure as employee complaints posted on social media can be collected and structured, or even quantified.

 

There’s a sudden increase in public sector data on the Chinese web after China began to promote the big data economy and the emerging artificial intelligence sector a few years ago. The Chinese government and public services of various sectors have been digitalizing their documents and other materials, to make them accessible to relevant third-parties or the public.

 

These publicly available digital channels are the main data sources for the newly emerged data vendors. Of the total data gathered by MioTech, a Hong Kong-based fintech developer that provides ESG data for finance and data analytics solutions, only about 10% are from companies’ voluntary disclosures. And with these non-traditional data sources, data vendors are able to track and assess ESG performance of private companies.

 

The latest technological advances like artificial intelligence came just in time to help make better use of the exponentially growing digital information and develop powerful analytics or monitoring solutions.

 

For instance, the major improvement made in the accuracy of object recognition has made satellite imagery useful to assess the environmental impacts of traditional polluting industries without having to depend on the reports released by regulatory inspectors, some of whom have themselves involved in corruption, and manage risks before these reports are released.

 

The predictive capabilities enhanced by machine learning can detect hidden vulnerabilities so as to flag up issues early on or send risk alerts based on news and posts on social media in the real-time or near real-time manner.

 

As the data-driven solutions are highly scalable, many data and analysis services providers are able to track thousands of Chinese companies simultaneously, or provide peer comparison or benchmarking functions and features. Data vendors like MioTech now cover almost all public and private Chinese companies.

Extracting information from the Chinese web, however, isn’t easy. Sophisticated technologies such as image recognition and natural language processing are applied for tasks that may sound trivial. For instance, in the public sector, neither are there national standards for document formats nor do public services or local governments intend to make information easily accessible to data collectors. As many documents are published in the format of JPG or PDF, technical techniques like image recognition are applied to extract information from them. With China being such a large country, with over 30 provinces and numerous local governments or public services who publish content in various formats, MioTech has some 12,000 ‘pipes’ going into different data sources to extract relevant information.

 

And it takes more data and efforts to make sure that proxy variables are reliable. To form a relatively accurate view from information gathered from social media, for example, much more types of topic data is needed, such as how long has the data been in the public domain, reposts on social media, the fade rate of specific information, according to MioTech.

Share this to
Share this to