With about 1,300 exchange-traded funds listed on Canadian exchanges, building an ETF portfolio is a project that can take hours and even days.
Or, you could get some help from artificial intelligence. My recent test run at ETF portfolio building using a pair of AI chatbots, ChatGPT and Chatsonic, was a qualified success. In seconds, both popped out built-to-order portfolios that mainly use popular funds you can trust.
AI doesn’t replace human intelligence – it’s too minimalist in providing the background supporting its investing choices, and some of its picks are too flakey. But AI has value to generate ideas for further research.
In case you’re not familiar with AI chatbots, they typically work as follows: you sign up for an account and then start a new chat by asking a question or making a request. For example, “build me a 70-30 portfolio of ETFs for a Canadian investor, with percentage weightings for each.” The blueprint here is 70 per cent stocks and 30 per cent bonds.
Here’s an AI irony: this cutting-edge technology has been programmed to bracket its financial information with old school cautionary boilerplate. ChatGPT actually mentions the idea of consulting a financial professional for personalized advice, and reminds you to conduct your own research and analysis.
The bit about advisers is silly – you don’t need them to invest successfully in ETFs. And conducting your own research is obvious. That’s because all you get from AI is a list of ETFs to consider – there’s not much additional information.
Here’s what ChatGPT produced in response to the request to build a 70-30 portfolio:
- iShares Core S&P/TSX Capped Composite Index ETF (XIC-T) – 35 per cent
- iShares Core MSCI All Country World ex Canada Index ETF (XAW-T) – 35 per cent
- BMO Aggregate Bond Index ETF (ZAG-T) – 15 per cent
- iShares Canadian Corporate Bond Index ETF (XCB-T) – 7.5 per cent
- Vanguard Canadian Short-Term Bond Index ETF (VSB-T) – 7.5 per cent
This is not a bad portfolio at all, though you could stop after the first three funds and be fine. XAW is a clever pick – it’s a global equity fund, which means you get exposure to U.S. stocks and markets outside North American in one package instead of having to buy multiple funds. Typically, investors buy separate U.S. and international equity funds.
Now for Chatsonic’s 70-30 ETF portfolio:
- iShares S&P/TSX 60 Index ETF (XIU-T) – 50 per cent
- Vanguard Total Stock Market ETF (VTI-A) – 20 per cent
- iShares MSCI EAFE ETF (EFA-A) – 10 per cent
- iShares Canadian Universe Bond ETF (XBB-T) – 10 per cent
- iShares Core U.S. Aggregate Bond ETF (AGG-A) – 10 per cent
The obvious glitch here is that bonds account for 20 per cent of the portfolio, not the requested 30 per cent. Also, a pair of U.S.-listed ETFs are included here, which means you’ll have to pay foreign exchange costs to convert your Canadian cash into U.S. dollars. Avoid that by finding a TSX-listed alternative to EFA and either going entirely with XBB or finding a TSX-listed U.S. bond fund.
If you’re using AI to generate investing ideas, try refining your questions and requests as you go along. Each tweak will change the results in interesting ways. For example, asking for a growth-oriented portfolio instead of a 70-30 portfolio prompted ChatGPT to produce a 10-fund ETF portfolio with exposure to clean energy, robotics, genomics and infrastructure funds. While this mix is excessively speculative, you do get insight into what sectors AI likes.
The best results from both AI chatbots came from requesting not a specific portfolio, but a list of ETF ideas. When asked for funds of interest to a Canadian growth investor, ChatGPT produced a list of conventional choices like XIC and ZCN for the Canadian market and the iShares Core S&P U.S. Total Market Index ETF (XUU-T) and the BMO Nasdaq 100 Equity Index ETF (ZQQ-T) for the U.S. market. One suspect pick was the Horizons Marijuana Life Sciences Index ETF (HMMJ-T), which was down 45 per cent for the 12 months to the end of May.
Other searches that produced useful results: the best ETFs for Canadian investors seeking dividend income and the best ETFs for seeking exposure to AI stocks.
In all of these lists, you can see both the strengths and weaknesses in AI-driven intel for investors. You get plenty of funds to research for yourself, but without much insight into why they were included in the first place. Chatsonic introduced its list of ETFs for dividend investors by noting that “these ETFs typically invest in companies that have a history of paying consistent dividends.” Notes for each fund basically paraphrased fund company marketing material.
Another issue with these AI-generated ETF investing ideas is that they focus strictly on the biggest, most established ETF players, and on traditional index-tracking funds as opposed to those with more elaborate or more niche strategies that could conceivably add value to a portfolio.
There’s some logic to this in that big funds tend to be heavily traded, which means you can buy and sell at competitive prices. Also, big index trackers are the cheapest to own in terms of their management and trading expense ratios, which is an important foundation for investing success.
There’s a growing number of good products available from smaller ETF companies, but they’re pretty much ignored by these two AI chatbots.
Are you a young Canadian with money on your mind? To set yourself up for success and steer clear of costly mistakes, listen to our award-winning Stress Test podcast.