Not logged in. Login | Signup

Artificial intelligence powered ETFs

2019-01-16, Michael Thompson

Share on facebook Share on twitter

Become an insider
Get the latest updates.

Artificial intelligence powered ETFs

ETFs that are actively managed by automated algorithms, such as the AI powered AIEQ, quickly gathered over $100M of capital after their birth in October of 2017. They were considered a sexy new investment vehicle in their early days but seem to have lost their appeal more recently. While several more years will be needed to judge their success, we’ll take a quick look at what they’ve done so far.

In short, these funds have mostly failed to justify their high fees. In its first full year of trading (2018), AIEQ lost about 10% including dividends. Compare that to a typical S&P 500 index fund like VOO which lost just over 6% by the same accounting. Even after including the sharp gains in late 2017 and early 2019, the fund overall is down since its inception (VOO is up slightly in the same period). The hope that these funds would better dodge market downturns has not panned out thus far. AIEQ has performed slightly better than the market in upswings but substantially worse than the S&P 500 in downturns.

BIKR is a quite different AI powered ETF. Using world economic data and models created by a team at Ocean Capital Advisors, it selects only international index funds (not individual stocks) and 1-3 year U.S. treasury bonds. I personally like this idea of selecting broader index funds (rather than individual stocks) for an AI driven model – it may lower the risk of limited or flawed logic in the algorithm. In the 6 months since introduction, it’s lost about 3.8% which is not too bad considering the markets in this period.

Personal Thoughts

While I am very pro AI (and computer automation in general), I have not invested in these funds. Trying to buy and sell stocks to beat the market consistently is complicated (and increasingly so with time). It’s not that it requires thousands of precise calculations (which a computer would excel at), but complicated in the sense that the strategies may be continuously evolving needing the type of generalization and adaptiveness that the human brain still triumphs in. The rare, good stock picks that do occur are often unique insights, not known / repeatable processes. On the flip side, I can easily envision an automated algorithm, with limited or flawed logic, making a few bad choices that drag its performance below the market. Consider Bridgewater Associates as an example. They’ve spent years recruiting extraordinary talent, purchasing massive computing power, and executing advanced algorithms. While they’ve had a few good years, their long-term performance has not been able to keep up with the S&P 500. Overall, I feel the risks outweigh the advantages. Couple this with much larger fees and I’m sticking with my simple, old-fashioned index funds for now.

Login to leave a comment.

Related Articles

Click here for a list of other recent articles.