5 ETFs and Funds Using AI for Stock Selection
A topic here on Nanalyze that has generated a great deal of interest is that of using artificial intelligence for stock trading. When we talk about stock trading though, it’s too vague unless we provide some sort of time frame. There are black box trading algorithms that can conduct 100s of trades per second and they don’t need AI to do that. On the other hand, there are AI algorithms being used by some of the world’s most successful hedge funds that capture alpha without anyone knowing what they actually get up to.
When it comes to your average retail investor, there aren’t many options for investment products that use AI for stock selection. We talked about this in an article titled Can Artificial Intelligence be Used For Stock Trading? Still, we can expect to see various investment products emerge that claim to use AI to generate alpha and retail investors will hop on for the ride. There are people out there who “invest” in ICOs proving that you can now sell people just about anything. Let’s take a look at 5 ETFs and funds that claim to be using AI to generate superior returns.
Horizons Active A.I. Global Equity ETF (TSE:MIND)
Launched just 6 days ago, this ETF is available to investors who can access the Toronto Stock Exchange (U.S. investors need an Interactive Brokers account in order to do this). The first thing that stands out here is the resemblance to a “fund of funds”. MIND charges a .55% management fee and then allocates their funds in at least 5 different other ETFs but no more than 20. As you might guess, the algorithms use more than 50 data points to choose which ETFs to allocate their funds into. It’s based on technology developed by a Korean firm called Qraft Technologies. An article by Bloomberg on the launch of this ETF quotes the co-chief executive officer, Steve Hawking, as saying the following:
I’m going to be buying some but I’m buying it as a nervous investor myself. We don’t know what the computer will do.
That’s quite the vote of confidence there Mr. Hawkings. We would hope that the allocation is being performed on ETFs that Horizon manages with fees waved otherwise investors are being hit twice with fees. While a fund-of-funds product offers extra diversification, the downside is that you’re paying more fees. There’s also the problem with any sort of alpha being diversified away. Then we need to consider transaction costs. How often does this thing get rebalanced? Are we telling the AI algorithms they can only rebalance once a month? This product is just not overly compelling from where we sit, but a closer look at the company behind it tells a much different story about a real commitment to using AI in the investment process.
Mirae Asset Management and Korean AI Funds
Most people who don’t work in finance (and some that do) may have never heard of Mirae Asset Group, a Korean firm with around $110 billion in asset under management (AUM). The Horizon ETF brand is owned by Mirae, and a closer look at what they’ve been up to with the Korean AI startup we mentioned earlier, Qraft Technologies, shows a number of AI powered ETFs that have been trading on the Korean exchange for much longer.
- Mirae Asset AI ASEAN Securities Investment Trust (1BS6133:KS)
Taken directly from Bloomberg, this is “an open-end fund incorporated in South Korea. The fund’s objective is to achieve capital gains by investing 60% or more of total assets to equities in ASEAN countries that are picked by artificial intelligence“.
- Mirae Asset AI Smart Beta Securities Master Investment Trust (1BK7832:KS)
If you are unfamiliar with the term “smart beta”, then be sure to read our article titled How Smart Beta ETFs of the Future Will Use AI. Here we see a product categorized as a “fund of funds” on Bloomberg which means it’s likely to choose among other funds and ETFs instead of individual securities. This goes back to our earlier concerns about using AI as a fund-of-funds manager.
When measuring the “success” of any investment product, you need to first pick a benchmark. According to an article in the Korea Times, they are using the standard benchmark for Korean stocks, the Korean stock exchange or KOSPI:
According to Zeroin, a fund analyzing firm, the six-month-old fund posted a 17.12 percent yield from Jan. 9 to June 23, well above the benchmark KOSPI’s 15.68 percent during the same period. In comparisons with one-month and three-month periods, the AI-powered fund showed higher yields than the KOSPI.
While that may have been the case at the time the article was written, a quick look shows that YTD returns for the KOSPI sit at +26.09%. Since the AI Smart Beta Fund began trading on January 10th, it has returned +23.26%. By that measure, it has underperformed the KOSPI and that sucks. An investment in a simple NASDAQ ETF (like QQQ) would have beat them both with a return of +29.88% and a whole lot less risk.
AI Powered Equity ETF (NYSEARCA:AIEQ)
The extent to which large media outlets butcher basic investing concepts is simply embarrassing, especially tech media. No, we’re not just talking about political commentary site Tech Crunch either. We’re talking about people who are supposed to be experts, yet dispense advice to the masses that is just plain wrong. For example, here’s an article from CNBC on the launch of this AI ETF:
The time frame they are talking about in that article is less than one week. That makes us think of all the armchair Twitter warriors who look at United Airlines stock falling -2.36% during one trading session and then say “United Airlines lost $415 million”. Or take the recent article from Tech Crunch that says some employee survey software called Glint will allow companies to “see a 40% increase in their stock price, compared to 4% for companies that only survey employees once a year“. That’s literally all they said. You have to put a time frame around these statements otherwise they are completely useless. And no, a time frame of a few days means nothing to your average retail investor. End of rant.
So according to an article published by MSN titled “Invest With An AI – An ETF Powered By An AI – AIEQ” (rolls eyes), there is now an ETF that you can invest in which uses
an AI AI to generate alpha. The firm behind the ETF is called EquBot, and to make things even more exciting, it uses IBM Watson to analyze millions of data points for a universe of 6,000 different stocks and ultimately holds between 30 to 70 positions. The expense fee is .75% and comes with a couple of expensive humans to oversee what the AI gets up to. According to an article by MarketWatch, “turnover is expected to be high” and the fund “could have a daily turnover of 2-3%” which could explain why the AI seems to have over-weighted firms like NASDAQ (second largest holding) and S&P Global (sixth largest holding). We see what you’ve done there you clever little AI algorithm. Of course high turnover = higher transaction costs = lower returns.
With about $73 million in assets right now, the fund seems to be popular with investors. We’ll definitely need more than a week to pass before we begin to see how it performs. With that sort of interest we’d expect to see more products launched so stay tuned.
BUZZ US Sentiment Leaders ETF (NYSEARCA:BUZ)
Since our last article on BUZ, they’ve had a name change to “BUZZ US Sentiment Leaders ETF” but still haven’t managed to attract many investors with AUM sitting at under $10 million. It could be their hefty .75% management fee, or the fact that their one-year return of +29.77% wasn’t able to beat the NASDAQ return of +36.67% over the same time frame:
The reason we used NASDAQ as a benchmark is because the ETF has such a high weighting in tech stocks. If we use the S&P 500 as a benchmark instead, we see BUZ coming out on top with a one-year return of +29.77% compared to an S&P 500 return of +24.24% over the same time frame. This is why it’s so important to choose the correct benchmark when analyzing performance. BUZ has been trading for about a year and a half now and the performance seems lackluster so far making us wonder just how long it’s going to continue trading.
We would expect to see more ETFs and funds popping up with all kinds of variations on the “AI stock selection” theme. If you know of any others out there, please drop us a line so we can keep a running list of them. We will continue to dig into investment products that claim to be offering something superior using AI when in fact they may not be. Lastly, it’s very clear that Mirae is committed to using Qraft’s AI technology throughout their investment products. We also see Qraft doing lots of work in the investment industry in addition to working with Mirae. Stay tuned for another article where we look more closely at Qraft. In the meantime, U.S. based investors will need an Interactive Brokers account to trade some of the products we’ve discussed.
Most people don't know that Interactive Brokers is the biggest online brokerage firm in the U.S. and the only firm we know of that allows you to buy stocks directly on 22 different foreign exchanges. We've used IB for many years as a cheap platform to exchange currency, for trading stocks on foreign exchanges, and for trading U.S. shares at just $1.00 per 100 shares. Click this link to get started.