Jul 24, 2017 05:18 PM

Bots or Humans – Who Is Better at Investing?

Toumi RA, a robo-adviser of Beijing-based fintech firm CreditEase, helped 99.6% of its clients make money in its first year of operation. Photo: CreditEase
Toumi RA, a robo-adviser of Beijing-based fintech firm CreditEase, helped 99.6% of its clients make money in its first year of operation. Photo: CreditEase

Liu Ze finally made his first profit from investing in stocks. And hats off to his financial adviser — Toumi RA, a robot.

Owned by Beijing-based online lender and investment platform CreditEase, Toumi had a track record that many human consultants would envy. Over the past year, it helped 99.6% of its customers achieve positive returns, much higher than the 10% typical of stock-market investors who did not consult the robo-adviser, CreditEase told Caixin.

Robo-advising and investing, which brings the expertise of pricey financial professionals to everyday investors, are still nascent in China. Toumi just celebrated its first birthday. Clipper Advisor, which markets itself as China’s first robo-adviser, was founded just two years ago.

Still, Tang Ning, CEO of CreditEase, has high hopes for the industry, banking on around 200 million active investors who do not have access to human advisers and asset managers because of their hefty fees.

“Our work is about teaching regular investors the concept of capital allocation,” Tang said, adding that many onshore investors are unfamiliar with the idea of an optimal investment portfolio based on one’s risk and return preferences.

Regulatory hurdles

In China, there are two separate licenses for investment advisory and asset management. Advisers make recommendations, while asset managers execute the actual trades.

This division of labor has created an existential problem for robo-advisers. Many of them work with financial advisory partners or get clients to execute their own trades to circumvent the difficulty of obtaining both licenses.

In March 2015, China’s securities regulator issued draft legislation on integrating oversight of the investment advisory and asset management fields, but there have not been any new developments since then.

Industry experts added that Chinese regulators have yet to formally define robo-advisers, let alone regulate them. This would slow, if not limit, the application of machine learning to financial services.

Bots or humans

For certain tasks, bots have already outperformed humans. San Francisco-based hedge fund Cerebellum Capital has recorded positive returns every year since 2009 by using algorithms that create and learn from their own trading strategies, while freeing up fund managers to do work that creates more value.

New York-based Rebellion Research has also replaced traditional stock pickers with a computer program that bases its recommendations on the trading histories of 54 countries over the last 20 years.

JPMorgan Chase & Co. has recently developed contract analysis software, called Contract Intelligence (COIN), which can in seconds complete work that used to take lawyers and loan officers 360,000 man-hours to do.

Asset management, theoretically, could be done without humans, said Zhang Jialin, chairman of fintech firm Zipeiyi Investment.

Zheng said there are seven stages in asset management: investor suitability analysis, asset allocation, portfolio composition, transaction execution, risk management, portfolio adjustment and post-investment analysis. Machine learning can be applied to each of these stages, he said.

Competitors or collaborators?

By 2025, 10% of the current jobs in the financial industry will be taken over by computers, management consultancy Opimas predicts.

For now, most artificial intelligence (AI) applications are still machine-specific. Robo-advisers, for instance, are data processors packaged and marketed as money managers. The software analyzes information provided by clients to determine their risk profile and financial goals. Clients are then assigned an investment portfolio based on algorithms designed by human investment professionals.

But the ultimate goal of artificial intelligence is to mimic the thought processes of humans, which scientists call “artificial general intelligence,” said Modar Alaoui, founder and CEO of Eyeris Technologies, which makes facial analytics and emotion recognition software.

These robo-advisers could one day determine a client’s risk level by tracking that person’s habits on the internet. The robot could then design a portfolio on its own by analyzing data and learning from its own investment mistakes.

Compared with humans, artificial intelligence (AI) needs less time to gain professional expertise. AI also processes data faster, does not make subjective judgements and is unaffected by emotion and other biases, said Liu Tie-Yan, principle researcher at Microsoft Research Asia.

“In the short term, AI is a helpful assistant; in the long term, AI is a potential competitor,” Liu said.

Having said that, he still believes humans are indispensable in finance. “The use of AI technology in the financial sector requires people with considerable experience to determine its framework, processes, basic logic and initial factors. These inputs must be adjusted according to market and industry developments,” said Li Hao, vice president of products at fintech firm WealthBetter.

Ma Yongan, founder of another robo-adviser Licaimofang, said without engineers and programmers, artificial intelligence cannot flourish.

“Artificial intelligence cannot create an investment framework; it can only optimize the existing system,” he said.

Contact reporter Liu Xiao (

You've accessed an article available only to subscribers
Share this article
Open WeChat and scan the QR code