Video Reports

Link to this video

Get LinkLicensePrint
Bookmark and Share

By Christian Charest | 11-20-2017

Horizons bets on technology with new ETFs

Steve Hawkins, president and co-CEO of Horizons ETFs, talks about a new fund managed by an AI system, and one that targets Robotics and Automation stocks.

Christian Charest: For Morningstar, I'm Christian Charest. The Horizons family of ETFs launched several new products during the month of November. I'm here today with Steve Hawkins, he is President and Co-CEO of Horizons ETFs Canada, to talk about two of those new funds in particular.

Steve, thank you very much for being with us today.

Steve Hawkins: Christian, thank you for having us.

Charest: Now, the first ETF I'd like to talk about is called Horizons Active A.I. Global Equity ETF, which trades under the ticker MIND. That's clever. It's the first equity ETF in the world to use artificial intelligence to make investment decisions. Can you explain to us how that works?

Hawkins: Sure. Well, a correction, it's not the first in the world, but just in the global equity allocation space, it is. There are single-name-picking AI systems and ETFs in the U.S. that are doing this type of programming, but nothing that has the global asset allocation strategy that we have provided here. And it is the first in Canada using the AI system.

To explain what the AI system is, it really is a neural network system. It's AI. It's computer-driven investment decision-making process. And it really – it picks an asset allocation from around the world. It's using country-specific or sector-specific asset allocations and then creating a portfolio from that. And the intent is to build a better portfolio than a human portfolio manager could and will generate returns that will outperform the global equity markets as a whole.

Charest: What are some of the parameters that it looks at to make those asset allocation decisions?

Hawkins: We looked at the global overall equity marketplace. We wanted to ensure that the AI system was reducing the overall risk of the portfolio relative to the global equity markets, as well as potentially outperforming the overall global equity markets. So, we limited the parameters to making sure that it invested a certain amount into North America, a certain amount into Asia, a certain amount into EAFE and Europe, just pluses and minuses relative to the world overall equity exposure from that perspective.

And we also took the standard deviation of returns of the overall global equity marketplace and we sort of reduced that so that it has -- AI does have the opportunity to invest some of the portfolio in cash if it feels that it wants to do that from a risk-control perspective. But there are limits and controls with respect to every single allocation that we input into the AI system for it to generate its outputs.

Charest: And how does it decide how much to invest in a specific country or a specific region?

Hawkins: Well, that's the interesting thing about the AI is we just don't know. So, at the end of the day, a portfolio manager could come to you and basically say, well, I'm investing 45% of the portfolio into the U.S. and this is why; this is the macro, these are the technicals, this is the global economic environment right now. The AI system is basically taking 10 years of back-tested history of returns of all of the different factors and there's hundreds of factors that it's inputting into its system for the purposes of determining at the end of the day what asset allocation it will make.

So, we've basically been teaching the computer for the past year how to become a portfolio manager, how to make the asset allocation for this ETF on a going forward basis. And every single time it goes through that decision-making process, and it can do many, many things that you and I can do in seconds that it would take us years to do from an investment decision-making, processing, strategy perspective.

The AI system ultimately will determine the portfolio and then basically, it's back-testing itself the whole time and it's saying, okay, now I've got data to determine what that answer was from and from there it's basically saying, now, I can make a better decision. Now, it's using all of that history and it does is over and over and over again, teaching itself. So, what decisions did I make which turned out correct, what decisions did I make which were not false but were negative from an overall perspective. And it's always working for the process of lowering the overall risk in the portfolio whereas outperforming the overall equity markets. And as long as it keeps doing things over and over again -- that's what an AI system does-- as it learns, it will ultimately make better decisions than a human portfolio manager.

Charest: So, are we talking about tweaking a strategic asset allocation, or does it try to anticipate the market?

Hawkins: Well, that's a very good question, because the AI system can look at the exact same data more than one time and it will never come out with the same decision at each time. So, we don't know what's driving the decisions, but the AI system is adaptive, and it learns from what it's done in the past and it will make a different decision almost every single time it's looking at the exact same data. Whereas, if you look at a smart beta or a factor-based index, it's going to look at the same data and it's always going to come out with the exact same output every single time. The AI system will always come out with a different asset allocation, but we just don't know why it's coming out with that asset allocation.

Charest: So, you mentioned that the program learns as it goes along. Are there plans to maybe broaden the set of decisions that it can make beyond regional asset allocation, maybe perhaps individual securities selection?

Hawkins: Yeah, absolutely. I mean, we see AI coming into our lives in many, many different aspects right now. I mean, portfolio managing is just the one that we are launching here with respect to this specific ETF and the global equity asset allocation. We see lots of different applications for this down the road with different types of other ETFs. This could be the next evolution instead of smart beta or factor-based investing. So, with factor-based investing, it's very, very one specific theme or a set of factors, but you are always going to come out with the same outcome.

With the AI system, it should always be endeavoring to outperform and beat itself. So, if it gets to repeat things and if it gets to learn, it should always be able to adapt and better itself. So, ultimately, this is the next evolution of factor-based investing.

Charest: The other ETF I wanted to talk about is called Horizons Robotics and Automation Index ETF, which you plan to launch at the end of the month. Now, this is a Canadian version of a fund that's been around in the U.S. for about four years now and has done very well, especially in the past two years. Now, the name of the fund already says a lot about it. But what can you tell us about its strategy?

Hawkins: Well, it's an index strategy, whereas the AI was an active strategy. Robotics, automation, AI from a theme perspective, which is the underlying for the index, has been the largest asset-gathering index-type strategy in the world this year. Robotics and automation are touching almost every part of our lives from our phones to our cars to what we eat to what gets manufactured to our healthcare system. It's in every single part of our life and it's only growing from that perspective. There's more and more applications for automation and robotics and AI in every single thing we do.

And just as a worldwide theme, we see more and more companies getting involved in this space, more and more companies developing robots. There's more companies -- you have self-driving cars these days. They are not just making the cars in the manufacturing warehouses and stuff, they are actually driving the cars for us now, right?

So, more and more automation, robotics being involved in everything that we see from a day-to-day perspective, everything we touch from a day-to-day perspective. And this theme has tremendous growth opportunities in the world. And this fund is investing not just in North America but in worldwide companies from a robotics, automation and AI perspective. And we just see tremendous growth opportunity in this space. There's nothing like this theme in Canada right now. And we created a business relationship with the ROBO Global guys and are bringing the ROBO ETF to Canada this year.

Charest: Now, despite the growth this is still a very specific niche of the market. How should the average investor use this ETF?

Hawkins: Well, it should be a part of their portfolio from a growth perspective. This is not a passive investment. But I don't look at it as a high-risk portfolio either, right? Almost 50% of the portfolio is in very well-established long-standing technology robotics, automation type of companies. And then there's another 50% of the portfolio, which is in, I would say, more earlier-stage but high-growth prospect companies. And the index really weights those companies very differently in the portfolio on an overall basis. But this is an investment for growth prospects. This is not an investment for the mom and pop who are trying to retire, and they just want to sit on a very passive equity or a balanced portfolio. This is a growth prospect for them.

Charest: Now, not to put you on the spot, but doesn't this bring back memories of another tech trend in the late 90s that also had tremendous growth for a few years; of course, I'm talking about the dotcoms. How is this different?

Hawkins: Well, I would say the single biggest difference is, this portfolio of companies that is being built inside this index and being invested into by the fund have long sustainable growth prospects. These are bricks and mortar companies. These are not companies that are just made in somebody's basement, created a website, started it up and somebody invested $100 million into them because they have an idea of what's going on. These are actual companies that are producing, that are making things, that are tangible that you're using in your everyday life and then you're only going to be using more in your everyday life on a going-forward basis. For every year to come we are going to see robotics and automation and AI be more involved in every single part of every person's life.

And from that perspective, the people that are building all of the infrastructure for us to be able to use from a technology perspective going forward, we are only going to see the growth in this industry grow astronomically.

Charest: And perhaps, the market has learned from its past mistakes?

Hawkins: I think it has. I think has.

Charest: Steve, thank you very much for explaining all this to us today.

Hawkins: Christian, thank you very much for having us here today.

Charest: For Morningstar, I'm Christian Charest. Thank you for watching.