Monday’s post discussed the inner workings of the two swap-based ETFs launched by Horizons in late 2010. Many investors are understandably concerned about the risks involved in these rather complicated funds, which get exposure to equity indexes without actually holding the underlying stocks.
I recently sat down with Jaime Purvis, Executive Vice-President, National Accounts, for Horizons ETFs, and asked him to explain more about how these products work.
Many investors get nervous when they hear the word “derivative.” Are we going to be reading in the next Michael Lewis book about how total return swaps blew up the ETF market?
JP: People think derivative is a dirty word. But a derivative is simply something that derives its price from an underlying asset. Your Michael Lewis reference is fascinating, because The Big Short describes how they parceled up all of these mortgages that people couldn’t afford, and they lumped together the good stuff and the bad stuff in all of these tranches and tiers. But the underlying assets were not very liquid. If you have a bad house in a bad neighborhood, it’s not like you can sell off a couple of rooms. And if the underlying asset is risky or illiquid, then so is the derivative.
However, when you look at something like HXT, which uses a swap on the S&P/TSX 60, you know there is liquidity there. Those stocks are in the index because of their size and their liquidity. So when you have an underlying asset like that, it’s a safe place to use this type of derivative, or swap structure. In fact, in pension plans across Canada, this is the primary method they use to get long-term equity exposure.
In the case of HXT, you are not paying a fee to National Bank, the counterparty. So what’s in it for them?
JP: Let’s say an investor gives us $100. We take that $100 and place it in a cash account that pays, say, 1.3% interest. Then we swap that return to National Bank, and what we get back is the total return of the S&P/TSX 60.
What’s important to understand is that the counterparty is always hedged. Ideally, if they are obligated to deliver to us the returns of the index, then they will hold those stocks in their own account so they are perfectly hedged. Then they’re essentially collecting a fee for doing the business. They’re getting that 1.3% and keeping that as their profit.
You’ve mentioned that banks are actually anxious for this business, so clearly they’ve identified another opportunity here—some kind of dividend tax arbitrage?
JP: That’s our understanding. There has to be some treatment of the dividends that is beneficial to them, or at least more beneficial to them than it would be to a Canadian retail investor.
[Dan Hallett of HighView Financial Group wrote an excellent blog post describing this idea in detail. According to Hallett, the bank “receives dividends which are effectively tax-free because they’re Canadian-source dividends. But the payment under the swap, which includes an amount equal to the dividends, is fully deductible. This tax arbitrage is effectively National Bank’s compensation for structuring the swap and taking on the tracking error risk.”]
How risky are these things?
Let’s talk about the risk of a swap structure. There seems to be a concern that if the counterparty were to default, then investors in HXT would lose everything.
JP: I think some people are confused about the difference between an exchange-traded fund and an exchange-traded note. With an ETN, you are absolutely tied 100% to the balance sheet of the note issuer. But a total return swap is structured so that what’s at risk is not the principal: it is only the gains, and only up to 10%. [Mutual funds in Canada—and that includes ETFs—are required to limit any derivative exposure to 10% of the fund’s assets.]
So let’s say you put $100 into HXT, and the next day the counterparty goes bankrupt. As a unitholder of HXT, you have no exposure to that counterparty, because you haven’t generated any gains, and they have no claim to your $100 investment.
So imagine I have $100 in HXT and this year the index goes up 9%, and now the value of my units is $109. Then the counterparty fails and cannot pay you that 9% return. You’ve still got my $100 in the cash account earning 1.3% interest, so now it’s $101.30. Does that mean all I lose is the other $7.70?
JP: Yes, that is the extent of your risk. And it’s important to note that any time the gains get to 10%, we would need to find an additional counterparty to take the further exposure, or we would need to crystallize some of those gains in order to bring that number down.
In any case, let’s envision a situation where one of Canada’s Big Six banks might default. The likelihood of one bank being in dire financial straits while the other five banks are not is remote—I think we can agree on that. So in any period when the counterparty might be in trouble, you would assume that the other banks would be down, too. In that case, given the role that Canadian banks play in the TSX 60, you could reasonably expect that the index itself would be negative. There would be no gains for the counterparty to deliver.
The worst case scenario, then, would be a single catastrophic bank failure during a robust bull market. And even then, that just applies to the gains that you have made. It has nothing to do with the principal.
Wouldn’t it make sense to have swap agreements with more than one counterparty, to spread the risk around?
JP: Absolutely. We’re in the process of assessing other counterparties now. It will depend on the amount of assets in the fund, and the maturity of the fund. Remember, it’s only been around since last September. We started with one counterparty, and we’ll look to add others.
I’m confused by the answer to the second question. Isn’t the 1.3% interest just compensation for the opportunity cost of the assets tied up being long the index stocks? This means that the 1.3% is not profit as stated. The only remaining incentive for the counterparty is the tax arbitrage explained by Hallett, which may, in fact, be a big incentive.
I’m confused as well. Claymore does the same with National Bank for CYH. But the conterparty has none of the holdings (CVY & HGI). The stocks are mostly mid-caps & not paying dividend.
@Michael: Great point. There does need to be a profit for the counterparty above and beyond the interest rate paid on the cash account. In most cases, there is an additional fee payable to the counterparty, but in the case of HXT, there is no fee because National Bank can profit from the dividend tax arbitrage.
@Eric: I will be writing more about the Claymore Advantaged ETFs soon. These do not use a swap structure. They use a “forward agreement,” which is rather different. Forward agreements always carry an additional fee (about 50 basis points), but the net benefit to the investor is a lower tax bill. More to come — I know this stuff is confusing!
My taxation is a bit rusty but let me try to clarify this a bit. Canadian corporations do have to pay tax on Canadian-source and eligible dividends. However, that tax is refunded once the recipient corporation, National Bank in this case, pays dividends to its shareholders. And since all banks, including National Bank, is already paying dividends, it effectively pays no tax on the dividends received from their long positions in stocks. So, it’s not a loophole per se because they would pay tax if they weren’t already paying dividends (if my understanding is correct). But it is a risk-free profit for them and that’s why they want this business.
Very good post and I like the comments Eric and Michael have made.
Does the hedge by National have to perfect? Can they hold two of the “better” (higher dividends, returns, etc) oil companies rather than all in the index for example? I’m sure glad the index owner is S&P and not National.
@Dan H: Thanks for the clarification. It’s interesting that this tax treatment isn’t better known. You blog was the only place I found it explained in detail.
@Steve: The counterparty is not obligated to be perfectly hedged: they are only obligated to deliver the index returns. It makes sense for them, of course, to hedge that obligation by owning the stocks in the index. They may well try to pick up an edge by overweighting some stocks, but if they underperform, they are on the hook to make up the shortfall. I’m pretty sure they would never do something as reckless as owning just two oil companies!
I am new to this area, and I apologize if these questions are naive:
1. Instead of being long on stocks, couldn’t National just hang onto the cash and directly earn 1.3% interest?
2. At first glance this all seems like a tax avoidance trick. Wouldn’t there also be legislative risk?
Very interesting, even if it’s still not quite crystal clear to me… But my understanding grows little by little.
I am looking forward to the information on Claymore’s Advantaged ETFs.
@DG: National does collect the interest on the cash. And in return, they have to deliver the return of the index (whether positive or negative) to Horizons. Since they know they will have to deliver these returns, it makes sense for them to simply buy all the stocks—then they know they will be able to meet their obligation.
Re: tax avoidance, it’s always possible that CRA could change its rules. But right now, as Dan Hallet’s post explains, no one is doing anything inappropriate.
@Lyne: These are pretty complicated structures that aren’t crystal clear to too many of us. I think it’s fair to say that if you feel uneasy about products like this, you should probably avoid them. There are more transparent alternatives.
Nice post CCP.
How often are the gains realized in this situation? weekly / monthly / yearly?
I would hate to be capped out during a good rebound year.
Also, which side of the agreement decides when to realize the gains?
@Chris: There is no danger of the returns being “capped out” because of the 10% limit. Given that a 10% gain in an equity index is going to be a regular occurrence, Horizons will have an arrangement in place to deal with it as necessary.
Ok, so National Bank holds the stocks, and the fund holds my cash. NB gets the interest on my cash, plus a tax credit, while I get the return on the stocks plus the dividends, and NB pays for any tracking error relative to the index.
Do I have that about right?
@Partick. You got it. The tracking error should always be 0.08%, which is the fee charged by Horizons.
Okay, so now that we have the basics of how the swap works, how does that play into the daily pricing of HXT? I’m assuming it just follows the TSX 60, but how does the absence of negative returns play into it (where NB doesn’t pay returns if the index goes below the starting point of the swap)?
@Chris: Swap-based ETFs move up and down with the indexes, just as conventional ETFs do. Although the ETF provider holds a cash account as collateral, it also holds a contract agreeing to swap this cash for the returns of the index, so that needs to be factored into the net asset value (NAV) of the fund. Think of it like your net worth: if you have $1,000 in cash but you also owe someone $200, your “NAV” is $800.
Thanks for the clarification.
I think these sorts of ETFs are allegorically like George Orwell’s Animal Farm. If you remember who was sitting at the table, at the end of the book, you would remember that you couldn’t differentiate between man and pig. These products don’t represent an indexed or passive strategy, any more than the pig at the end of Orwell’s masterpiece still represented a pig.
Over the long run, only the brokers will really benefit from products such as these. Most people who do well in the markets do so because they keep things simple. If you let something tickle your gambling bone, you could do well for a while, but I don’t think you’ll do well for long–at least when compared to a simple, indexed portfolio over a lifetime. It pays not to think too much, when investing.
@Andrew: Thanks for the comment. I appreciate simplicity, too, and I don’t think anyone should use products like this if they don’t understand them. But they’re not speculative. They do indeed represent a passive strategy: they’re designed to track an index and nothing more. They just do it in an unconventional way for tax reasons.
Most people may not realize that RBC, Scotia and Altamira index funds all use derivatives to get exposure to US and international stocks. I believe this is a holdover from the days when foreign content was limited in RRSPs. But the idea was the same: get exposure to the markets cost-effectively while still adhering to a government requirement.
I’m not smart enough to understand these, I’ll admit. But I’ll do OK without them, I think. If I can’t explain it to a fifth grader…. That’s my motto.
I do trust you, Dan, when suggesting that these aren’t a step towards the active management concept of decision-making. I’ll admit, I understand these products so little, that I couldn’t give an intelligent opinion on them. I suppose I’m now like the reluctant investor you and I spoke about: the person first introduced to an indexed strategy, but reluctant to change from an actively managed strategy.
That said, I own my Vanguard first world international index, my Vanguard total U.S. market index and my Canadian bond index, and (stubbornly perhaps) won’t budge. I think that new products will always become available. And I think that most investors will spend time researching those new options and shifting their money to the next best thing. Do you think the quest for a perfect strategy could become the enemy to an investor’s eventual, long term returns?
@Andrew: “Do you think the quest for a perfect strategy could become the enemy to an investor’s eventual, long term returns?” Absolutely. I think we all fight with this occasionally — we worry about which products are the best, which asset allocation is ideal, etc. These are important issues, for sure, but there is no perfect solution. There are lots of excellent, if imperfect, solutions. If we all just picked one and stuck to it we’d do fine!
I’m probably the slow kid in the class, so bear with me on the slow kid kind of questions.
If, say, National Bank really held those stocks within the index, wouldn’t they incur costs related to bid/ask spreads and commission charges to buy/sell those stocks? And in that case, they’d be earning slightly less than the index’s return, wouldn’t they? So if the total return of that index was 8 percent, and the National Bank was on the hook for 8 percent to Horizons, then wouldn’t they have to pay Horizons more than they earned through holding those stocks, if they had to pay the index’s absolute return, before costs? What remains, for National Bank, is the paltry 1.3 percent they’d make in interest, which would have to be widdled down to 1.2 percent after expenses associated with owning those stocks. I understand that owning a small number of stocks (in such a small index) would not be as expensive as a very large, broad index, but even a non-profit company, like Vanguard, charges expense ratios of roughly 0.1 percent, so we’d assume that National Bank would incur costs of at least this.
If that’s the case, I’d be critical of owning National Bank shares, as it doesn’t sound like a great business arrangement for them.
What am I missing?
@Andrew: as I understand it, the counterparty can make money in three ways, and the 1.3% interest is not one of them. (They could earn that themselves, after all.)
1. They get paid a swap fee by the ETF provider. That is the case in almost all swap-based ETFs. HXT is the exception, because it uses reason #2.
2. With HXT, the bank takes advantage of a tax arbitrage opportunity. The return they earn on their exposure to the TSX 60 is a little higher than what they are obligated to deliver to Horizons, and they pocket the difference.
3. They choose to take risks with the money they get from Horizons by investing in something other than the TSX 60. If they earn excess return, they keep it. However, if they lag the index, then they’re still on the hook to deliver the index returns. So this risk is not passed on the ETF investor.
I’m comfortable with Horizon’s swap arrangements. I’m not sure that all other swap-based ETFs have such strong risk-mitigation measures for the risk of counterparty default.
I would like to know when horizon actually realizes the gain. Is it annually? If you start with $100 that horizon is holding in cash, and the return for the year is 10 percent, when does horizons actually start holding $110 in cash. I’m trying to figure out the risk of holding the ETF over a course of say 10 years. Are you at risk of only losing the current years returns if the counterparty defaults?
@KW: Horizons always holds cash, and the amount held is adjusted continuously as the value of the fund increases. Horizons is legally bound to make sure the counterparty risk never exceeds 10%, so the length of time you hold the ETF is irrelevant.
Dan, now that it is more than 2 years since these two ETFs (HXT and HXS), are you able to do an update regarding their performance, modifications if any, and their popularity?
It will be interesting to find out how results and opinions evolved over these period.
@Be-en: The only change has been that HXS has dropped the currency hedging, which makes it a better product, in my opinion. Both funds have tracked their indexes as well as one would expect after fees. Because of its extremely low fee, HXT has done a particularly good job in that respect. I have few concerns about the counterparty risk in these ETFs, and I use them myself.
How does the National Bank profit? If they have to invest their own capital, why would they return the gains to Horizon for a paltry interest fee?
@Vreni: It may not be a large fee, but it’s a risk-free return for National Bank. There also seems to be some tax arbitrage:
@Canadian Couch Potato:
Based on the information you have provided in your blogs/articles, I understand the risks involved with the swap agreement. Thanks! I understand those risks are minimal for several reasons (only 10% max, only to gains, etc.).
However, the risk which concerns me, which I have not seen addressed, is the cash account risk. If I understand correctly, Horizon deposits the ETF assets, ~ $1.1 billion, in a cash bank account. This is basically an unsecured loan to a single bank in an amount far exceeding CDIC insurance. If that bank defaults, 100% of the swap-based ETF value is lost. The swap-based ETF investors hold no hard assets, whereas a physical ETF holds actual company shares.
I look forward to your thoughts on this …
@Brett: I would expect that Horizons holds Treasury bills with the cash.
Need your guidance to build tax efficient etf portolio for non-registered account.
I have been using HXT and HXS for about four years……..I am very pleased with both, and I no concern about National Bank counterparty risk.
They are perfect for tax efficiency in a non-registered account, and I even have some in registered accounts due to the low mers and tracking errors.
My only concern would be CRA changing the rules at some point…..would that mean investors would need to sell the shares, or would Horizons be able to change the mandate without sales?
@Ken: The most recent federal budget has hinted that the government may indeed change the rules. But at this point, no one knows exactly how the situation will play out:
Dan, just wondering why you haven’t ever used these swap ETFs in your model portfolios. Specifically, if you were to make a model ETF portfolio today that had an ETF for Canadian Equities, would you use HXCN? Why not?
@Evan: This will offer some perspective: