Artificial Intelligence Investment Management?


#1

Are there any AI investment managers out there?

It seems to be like AI could actually mitigate some of the issues around markets (e.g. panic selling, holding equities long-term…)


#2

I know there are definitely investment products which use use to track other stocks. IBM’s Watson is being used to full effect in the indexing business and the ETF business as well…

I think what you are interested in something which gives genuine advice to its users, right?

These systems are definitely in their early days, we haven’t got there yet. When they do come in, I think it’s very unlikely we will see any substantial change at all in the market. For a short period of time one fund might have an advantage, but then things will catch up as they always do.

If a hedge, or company has an informational advantage (a special news source, an algorithm, a specific strategy), their competitive advantage will not last long. A great example of this is in the Michael Lewis book, Moneyball.

link: https://en.wikipedia.org/wiki/Moneyball_(film)

In the book, the main character realises he can use statistics to great the perfect baseball team from a team of misfits. His tactics worked incredibly well and his team succeeded far beyond what anyone expected. When rival teams discover his approach, they copied it. Once everyone in the league adopted his strategy, his competitive advantage was eroded.

You see a similar thing with college endowments. Yale’s college endowment fund significantly outperformed its rivals in the 1980s, as it moved into “alternative assets” that other institutions ignored (PE, leveraged buyouts, venture capital). Fast forward forty years, everyone is using the same strategy and it is no longer effective…

link: http://fortune.com/college-endowments-investing/

My point is… investing in a fund uses “artificial intelligence investment management” is unlikely to have superior returns for a very long. An algorithm is nothing more than a computerised, automated strategy which can be broken down and copied.

When a strategy is successful, people copy will copy it. And it will cease to be successful.


#3

Wow, thanks for that. Really interesting!