The authors of the paper summarized below admit their analysis could be better. It could have had a wider scope and it would have been better if it’d been conducted over a longer time period; but, as their subject didn’t exist until October 2019 I think we can forgive them. If their work has Swiss-cheese-like holes, at least the space between them is filled with cheese.
Yue Guo (et al.) from the Southern University of Science and Technology wanted to see if the use of an AI investment assistant (hereafter AI-analyst) changed investor behavior. If it did, in what ways was it doing this and did investors end up making more money as a result of the use?
On October 1st 2019 an AI-analyst function ‘Zhi Xiaobao’ (支小宝) was introduced on the Ant Fortune platform which was, and still is, the largest retail investment platform in China. Initially with only limited functionality the feature developed to allow Q+A interrogation along with its other data gathering/supply-to-customer capabilities.
As there were before and after investors on the platform the researchers could analyze what changes occurred as users started to click the little blue ant which was now appearing on their screens, top left.
They discovered as follows:
- AI-analyst users performed more manual investor sessions and invested more than those who didn’t use the tool
- Risk tolerance of users increased i.e. they traded more products with increased volatility characteristics
- Return results seem inconclusive (IMHO, the paper is differently nuanced). Some investors appeared to do better, some a little worse
Not in the paper but my two-pennyworth would be to note, from the platform operator’s perspective the introduction of an AI-analyst appears to have been an unequivocal success. Clients ended up trading more and in more products. Huzzah, for the operators at least.
However, as seasoned investors know, and there’s a body of literature spanning decades that backs the notion up, one of the biggest enemies of long term performance is over-trading and the transaction costs associated with the behavior.
Those who preach the benefits of greater ‘financial inclusion’ (the term comes up 17-times in the paper) are, I notice more often than not, partisans recommending behavior which surely benefits them but rarely, if ever, makes sense for investors. Where are the client’s yachts, and all that?
You can access the paper in full via the following link AI and Investor Activity.
Happy Sunday.