Market-Linked GICs: Not Now, Not Ever
Markus Muhs - Aug 19, 2021
A recent advertisement about "stock market GICs" has me again going on a rant about how terrible these types of investments are, so I dug up and updated this old blog post.
Market-Linked GICs (or "stock market GICs") are fairly opaque bank products, geared toward customers not willing to risk losses, but also not happy with standard GIC rates.
They are complex, and often their own salespeople don't even understand them.
In both good times and bad times, ML-GICs can underperform supposed riskier investments.
Dividends are an important factor in long term investment returns, and ML-GICs ignore them in their return calculations. They are also very tax-inefficient.
ML-GICs benefit the banks and their salespeople, not you.
Recently, I got the following in my e-mail in-box. The institution responsible shall remain unnamed, but it sparked me to look up this old blog post and update it, to help inform the consumer who might fall for such things...
Stay tuned. I think I'll update my blog post about how crappy of an investment these are... pic.twitter.com/iJrsW9ydFV— Markus Muhs, CFP®, CIM® (@CGWM_Muhs) August 19, 2021
First, an admission: in my career, I actually sold these types of GICs when I was with a major bank. I used to even think they were a good idea. In fact, when I first learned about them during my training at the bank’s call center (where, believe it or not, call center phone operators with zero financial background were able to advise customers on investments and recommend these GICs) I thought “these are genius; why does anyone even buy mutual funds anymore?”
If you have never heard about ML-GICs before and have no idea what I’m talking about, consider yourself lucky. You still might want to continue reading anyway, just in case at some point in the future some bank pitches such a product to you.
ML-GICs come in various forms, with a huge variety of rules as to how their returns are calculated. Like any other GIC they are 100% guaranteed by the issuing bank, eligible for CDIC coverage, and nominally cannot lose you money. Whereas a normal GIC pays a set rate of interest for a term, a ML-GIC pays out a variable interest that depends entirely or in part on the performance of a particular stock index or combination of indices.
Some ML-GICs “cap” returns to a certain maximum, while others use a percentage multiplier to reduce your potential upside return. To the average consumer, this seems fair, because you would expect to give up some upside potential if you’re eliminating the potential for losses.
The simplest ML-GICs are calculated based on the value of an index at the start date of the term and the end date of the term, then capped if that return exceeded the maximum. For example, using approximate rates that were available back in the day (today they are much lower), you could buy a GIC linked to the Canadian bank index with a maximum return of 50% over 5 years. If you happened to purchase this GIC in 2004 and it matured at the beginning of March of 2009 (market lows), you made 0% and got your principal back. If you bought in 2009 and it matured in 2014 (during which the index went up over 150%) you got the maximum 50% return.
Seems like a good deal, right? 50% over 5 years is 10% per year, much higher than you can get on a regular GIC (actually, broken down as a compound annual growth rate it works out to 8.45% per year) and there’s no downside, whereas I would have lost a lot of my investment in 2009 if I had it in mutual funds, right?
Well… yes and no on the last one. Rather than compare with an index (which nobody was buying back then), I'll compare with a likely alternative investment you would have considered at the time; the same bank's popular top-selling Dividend Growth fund (particularly heavily invested in Canadian banks). It did lose quite a bit during the financial crisis at the end of the term. From start (in the case of the above: Feb 29, 2004) to end (Feb 28 2009), probably the worst 5-year span for financial companies in our lifetimes, the mutual fund however squeezed out a small positive return of 1.08%.
Over the following 5 years, during which the Banks GIC paid out its maximum 50%, the mutual fund returned a total of 109% net of fees, or 15.92% annually.
So, in this example, the purchaser of the ML-GIC saw no benefit during the really bad markets (he made 0% when he could have gained 1.08%), and the trade-off was giving up 7.47% annually in returns during a good period. Not even the most overpriced commission-laden mutual funds could have done worse!
The Importance of Dividends
One similarity among all (or at least most I’m aware of) ML-GICs is that their market return calculations are based entirely on differences in index level from the start of the term to the end, and do not in any way account for dividends. This is the biggest major strike against them, as generally over the long term dividends account for a significant portion of an index’s total return. While they are often marketed as getting partial market upside, with no downside, what the banks define as "market upside" is practically irrelevant to what these indices actually return.
The above graph compares the TSX Composite Index that we usually see quoted in the news (red) with the TSX Composite Total Return with dividends reinvested (blue) over the 10 years from the 2008 highs to 2018. The TSX didn't really go up much in that period, from the 14,000s in 2008 to the 16,000s in 2018.
That pathetic 1.14% annual principal growth rate, however, is nicely enhanced with dividends to get up to a halfway decent 4.19%. On a $100K investment, it’s the difference between ending up with $112K vs $150K. The former is what a typical ML-GIC bases its return on!
This issue with dividends compounds by the fact that often the ML-GICs that are promoted the most by the major banks are linked to indices that generally earn half or more of their total returns from dividends, such as financials and utilities. The banks will steer customers towards such with promises of higher maximum returns than on an S&P 500-linked GIC, for example, which generally earns a lower dividend yield.
While the previous example follows a very simple start point/end point index level calculation, capped to a maximum return, here’s an example of how a different bank calculates returns, and just how convoluted the "rules" can be. With some of them, I wonder if the average salesperson at the bank can understand them, let alone the average customer.
The Rate of Return for the Term will be the return on the best performing Reference Portfolio during the Term multiplied by the Participation Rate, except that the Rate of Return for the Term will not be less than zero. The return on a Reference Portfolio is equal to the weighted average of the percentage changes in the values of the indices comprising the Reference Portfolio, calculated as a percentage using the weightings set out above. (and it goes on…)
I actually went through the “rules” in the fine print and calculated the likely returns an S&P500-linked GIC would have made during a really bad 5-year span (Oct 1st 2007 to Oct 1st 2012) and then what it would have done over a 5-year span of more normal returns (Oct 1st 2008 to Oct 1st 2013) and compared both to investing in SPY (the SPDR S&P 500 ETF), adjusted to Canadian dollars.
On the GIC, the rules are as follows:
- 5 year term.
- Opening market level is the index level the day of purchase.
- Closing market level is the average of the final 12 months of the term. They either do this to avoid the bad luck for the client of having the GIC mature on a bad day, or the bad luck for the bank, of it maturing on a particularly good day.
- Participation rate of 45%. Final return calculation is multiplied by 45%.
It's easy enough to calculate the return of the GIC from 2007 to 2012:
- Oct 1st, 2007 index level of S&P 500: 1547
- Oct 1st, 2012 index level of S&P 500: 1444
- Return = 0%
SPY's total return in $CAD over this time span:
Annualized, this comes out to 0.71% and serves as a nice reminder to market-timers that simply staying invested during this super-volatile 5 year time period still yielded better than a money market fund. Remember also these are the returns in $CAD, and this demonstrates the benefit of not using a 100% CAD hedge on your U.S. investments: you miss out on significant downside protection (the actual SPY in USD returned 0.6% annually over this time period).
So in a really crappy time span, including the worst bear market in 80 years, the ML-GIC investor was worse off!
Now, let's look Oct 1st 2008 to Oct 1st 2013, which I chose not because it's the best 5 year span, but because it's a 5-year span during which SPY returns were a fairly average (by long-term standards) 9.3% annually (again, translated to $CAD). Here's where calculating the returns of the ML-GIC gets a lot trickier using the rules from above:
- Oct 1, 2008 index level of S&P 500: 1161
- Oct 1, 2013 index level of S&P 500: 1695
- Average closing value of the 12 months leading up to maturity: 1564
- Gross index level return: (1564-1161)/1161 = 34.7%
- Times 45% participation rate: 15.62% total return
- Annualized over 5 years: 2.95%/a
So, to sum up all of the above: In a really bad 5-year span the ML-GIC investor gave up 0.7% annually to avoid losses, and in an average 5-year span he gave up over 6% annually!
Only in a very rare 5-year span during which the markets are down more than the cumulative yield of the index's dividends, would the ML-GIC investor will finish better off. If that kind of risk is an issue–if the investor's time horizon is only 5 years–they should be looking at a fairly conservative balanced portfolio anyway, which also should yield better outcomes than the GIC.
I could go on and on about why ML-GICs are to be avoided, but to keep this post short(ish), I'll finish with one additional reason: taxes.
In a registered plan, of course, how inefficiently these investments are taxed doesn't matter. In a taxable account though, you should simply never own them.
On a mutual fund or ETF you're taxed annually on a variety of investment incomes: capital gains, interest, and dividends. Some of this income is taxed fully at your marginal tax rate, and some (such as capital gains and Canadian eligible dividends) at a lower rate. Importantly, a 5-year investment in a mutual fund or ETF generally spreads the taxes over 5 years, with the exception of deferred capital gains when you sell. Presuming you don't sell your portfolio after 5 years though, much of this can continue to be deferred over the long term.
On a 5 year ML-GIC, not only is all income taxed at your highest marginal tax rate, but it's also all taxed in one year–the year in which the GIC matures.
With most GICs–where a fixed interest rate is calculated and compounded annually–even if the investor doesn't receive the interest until maturity, the taxes are still paid annually. With a ML-GIC there's no way of knowing how much interest will be paid until the end of the term.
Conclusion: Why do the banks like 'em?
The banks have continuously tinkered with ML-GIC products over the years, trying to get the perfect balance of a profitable product that's just opaque enough to attract the average investor and for their sales reps to not second-guess.
As I mentioned earlier, when I first started at a bank I thought these products were great. My understanding was that they offered mutual fund-like returns–up to a prescribed maximum–with a backstop. Shows how little i knew, as a seller of them. At that time the bank was also still figuring things out. For example, back then they offered a GIC linked to the TSX Energy sector which offered a 100% maximum return. People who bought it in 2003 in their RSP were delighted to see it double upon maturity. The bank didn't offer it anymore after that.
As it stands today, most of these ML-GICs offer fairly low maximum returns, made attractive-looking to the unsuspecting retail investor by the fact that they are almost always quoted as 5-year total rates of return, not annual percentage rates. "Potential 20% return" on a 5-year GIC looks great to someone who doesn't realize it's a total return over 5 years, and if you don't know how to calculate an annual compound rate from a total return, it still looks like a nice 4% annual return (when it's actually 3.7% per annum).
The latest GIC ad that I made reference to in the tweet at the top of this post promised in the e-mail, on their website, and on their "info sheet" a 12.7% total return over 5 years on an S&P 500 linked GIC, with an asterisk. Absolutely nowhere could I find the explanation that this was not the annual return, just a "conditions apply" in the fine print; probably an explanation when you go in-store. Deceptive advertising. That 12.7% total return is a mere 2.42% annual compound return, which is just a tiny notch higher than the 2-ish% you can get on a standard 5-year GIC in 2021.
For the bank, this is a pretty sweet deal. Worst case (for them), they end up paying a bit more than they otherwise would have on a regular 5-year GIC. For this outcome, I believe they buy some form of insurance in the background. In the best-case outcome for them, they borrow money off you for 5 years and pay you very little–and you're totally fine with it because you didn't lose! While most of these GICs used to pay 0% if the market levels went down, they tend to offer a tiny minimum return, as you can see in the ad at the top of this post.
As a former salesperson at the bank, these GICs were also an easy way out of two typical customer conversations:
- Instead of bartering on GIC rates, or losing most or all of my "sales revenue" bonusing the rate on a traditional GIC, I could simply offer the ML-GIC which is both more interesting than a "vanilla GIC" and the customer felt they were getting a better deal (and I got full sales revenue).
- Instead of going through all the trouble of opening a mutual fund account and putting a customer through a risk questionnaire, I could simply click a few buttons and plop their money into a ML-GIC that paid me just as much sales revenue as the mutual fund. On a busy RRSP deadline day this was crucial in order to hammer out as many sales in the day that I could. For less experienced sales reps, or those just not comfortable with mutual funds, this was also an easy out.
Of course, it didn't take me long–even in the bank salesperson role–to come to the realization of what a bad deal these are. When I was a financial planner at the bank, and sales reps referred clients to me, I explicitly warned them about ML-GICs; that if they refer clients to me who own them, and the clients ask questions about them, I'll answer truthfully and it might not make the sales rep look good.
If you have any questions, don't hesitate to drop me a line.
Markus Muhs, CFP, CIM
Investment Advisor & Portfolio Manager
Investment Advisor & Portfolio Manager
Never miss an update! Sign up to my eNewsletter
Graphs sourced from Morningstar Advisor Workstation. Image of coins is my own.