Asset management firm of the year: Ping An of China Asset Management (Hong Kong)

Asia Risk Awards 2020

Asia Risk Awards 2020

Ping An of China Asset Management (Hong Kong), a wholly owned subsidiary of China’s largest insurer, is named asset management firm of the year for its efforts to put technology at the core of its business, not just in order to be able to capture alpha more effectively but also to help identify looming risks and promote diversification within portfolios.

“We are not here simply to have the highest absolute return, but to try to maximise risk-adjusted returns,” says Chi Kit Chai, head of capital markets and chief investment officer at the firm.

Chai joined in July 2017, after 21 years at US pension fund Teachers Retirement System of Texas. He has worked hard to promote the use of technology within its investment decisions.

“The investment management industry is slow in terms of using AI technology,” says Chai. However, he says Ping An aims to build the next generation of investment technology, so when it started in Hong Kong, it decided to borrow the AI experience and expertise from its parent company.

Ping An Asset Management started to develop its AI platform back in 2018. With the technology resources from its parent company – where AI has been widely used in its insurance and banking services – the offshore asset management firm brought it into the asset management business. In building the new technology system, the firm recruited from both the investment world and academia to form the backbone of the team.

Ping An’s AI system is inspired by Google’s AlphaGo artificial intelligence program. It employs deep-reinforcement learning techniques – which improve over time through a ‘trial and error’ approach – to fine-tune its investment strategy.

Ping An Asset Management uses both historical structured and unstructured data – including natural language news, price movement information, macroeconomic inputs and company-specific accounting information – in order to train its AI algorithms.

Chi Kit Chai
Chi Kit Chai, Ping An

The AI is currently monitoring more than 300 factors and selects between 20 and 50 factors to construct its portfolios every month.

“This platform enables us to generate better information ratios,” says Chai, referring to the widely used measurement that helps investors understand their return over and above the benchmark.

He says Ping An’s information ratio is currently around 4, noting that anything above 1 “is considered very good”. This achievement can be attributed to AI, he adds.

“It would be nearly impossible for human analysts to do such a large amount of data analysis,” he says. “With AI, asset managers can gain a better understanding of the risks they are taking and where the future alphas lie, so that they can maximise risk-adjusted returns and build their portfolios accordingly.”

This is particularly important in publicly traded stocks, continues Chai, where the markets tend to be fairly efficient.

“If you want to try to generate alphas in the public markets you have to have an edge. Technology is that edge,” he says. 

In 2019, Ping An launched two AI-driven multi-factor funds on Chinese A-Share equities based on this new technology platform.

Each portfolio consists of around 100 stocks. Chai says the portfolios rebalance on a monthly basis, using AI to identify the favoured factors of the moment. The turnover ratio – the measure of how quickly stocks are replaced in the fund – has been constrained to be no more than 20%, thereby reducing costs for investors and reducing market impact.

Putting together the AI platform has not been easy, though. Despite AI being such a hot topic in finance at the moment, it’s very difficult to find AI engineers who want to join an asset management firm.

“They’re more attracted to fields like autonomous vehicles – many of them are not as interested in working in the finance and investment industry,” says Chai.

The linear nature of financial data also makes it difficult to combine AI technology with asset management.

“Financial history is a single path of time-series data. That makes it more difficult to do statistical analysis as the signal-to-noise ratio is low and results are more prone to be spurious,” says Chai.

Chai contrasts the nature of financial data to other fields such as facial recognition in which AI techniques are deployed and engineers have a lot of properly labelled data to train their models repeatedly.

For example, if a facial-recognition system was fed with enough pictures of cats, then it could easily be taught to accurately identify a cat in the future, because, although every cat is slightly different, they each have enough in common as a species to be uniquely recogniseable. But for finance, you simply cannot do that as you only have one single path and the market is different every day, says Chai.

The answer lies in non-linear algorithms, which process a large amount of data in order to infer potential underlying relationships between different elements in the market, and therefore to drive investment predictions.

Linear analysis, where different data points can fit into a straight line curve, has its limits “mainly because the world is non-linear”, says Chai.

With its roots in China, Ping An Asset Management is now expanding its capabilities overseas.

One client – the managing director of a British fund that invests in Chinese equities – says: “They’re committed to generating alpha and active fund management, which coincides with what our clients are looking for in the region. Being a Chinese firm, the insight they bring to China equities is invaluable and give us an edge in the market.”

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact info@risk.net to find out more.

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here