Some of my readers ask why I bother with ETFs when I could save the MER costs and buy the index stocks directly. For example, instead of owning the popular iShares index fund XIU, why not save the 0.18% per year and buy the 60 Canadian stocks that make up the index this ETF tracks? The short answer is that replicating XIU cheaply is difficult.
Let’s look at a fictitious example. Dan has a sizable portfolio. His allocation to XIU is $500,000. This fund’s MER of 0.18% costs Dan $900 per year. With such a large portfolio, Dan seems like a good candidate to try buying the TSX 60 stocks directly to save some of that $900.
Let’s assume that Dan’s XIU shares are all held in tax-sheltered accounts, such as RRSPs, RRIFs, or TFSAs, so that we don’t have to worry about capital gains taxes. We’ll also assume that Dan pays only $5 per trade and that the spreads on TSX 60 stocks average one cent lost per $50 traded (0.02% per trade).
Trading Costs
If Dan wants to own all 60 stocks in exactly the right proportions, he will have to make 60 trades (not counting the initial sale of XIU). This will cost Dan $300 in commissions and another $100 in spreads. Even if Dan has his money split across multiple tax-advantaged accounts, he can just own a different subset of the TSX 60 in each account. So, the initial cost is $400.
Whenever Dan adds new money to his accounts he will have to make 60 trades if he wants to maintain the correct proportion of each of the 60 stocks. This is another $300 in commissions and a small amount in spread costs (assuming the deposit is relatively small).
Dan will incur the same $300 in commission costs whenever he makes a withdrawal. There is the added complication that the account he is withdrawing from may not have any shares in one of the stocks. This will force Dan to sell more of a different stock and then make two trades in a different account to get back in balance. So, withdrawals could easily cost Dan closer to $400 in commissions.
Depending on how often Dan makes contributions or withdrawals, trading costs might add up to more than the $900 per year Dan is trying to save.
Index Changes
Another source of costs for Dan is changes to the TSX 60 index. For example on 2013 June 24, the weight of Valeant Pharmaceuticals increased in the TSX 60 index. To account for this, Dan either has to add new money to his accounts to buy more Valeant, or he has to sell some of each of the other 59 stocks to buy more Valeant. The cost for this is about $300.
Rebalancing
Dan also owns some bonds and some U.S. stocks along with his Canadian stocks. He has target allocation percentages for each of these three categories. He finds that he doesn’t have to rebalance often, but when he does, he gets hit with another $300 in commissions for changing his allocation to Canadian stocks.
Being Flexible about Following the Index Exactly to Save on Costs
An obvious way for Dan to save on costs is to be a little flexible about owning the 60 stocks in exactly the correct proportions. The problem is that this has a cost, even if this cost is not visible.
For example, Dan might choose to own only the 20 biggest stocks instead of all 60. This will instantly cut his trading costs by two-thirds. Unfortunately, this will slightly increase the volatility of Dan’s portfolio which leads to volatility losses. This might seem like a small effect, and it is, but we are comparing it to a small MER of 0.18%. Under one set of assumptions, the volatility losses amount to 0.06%, or about $300 per year.
If Dan is clever, he could examine all these factors and find an optimal strategy that minimizes the total of all these costs (commissions, spreads, index changes, rebalancing, and increased volatility drag). This strategy would involve following the index fairly closely, but not so closely that Dan has to make a lot of trades each year.
Let’s suppose that Dan manages to find a strategy that is expected to cost him only $700 per year instead of the $900 XIU is charging him. Exactly how much work should Dan do to save this $200 per year? Personally, I’d much rather just buy XIU and spend more time with my family and friends.
On the subject of what I like to buy, I like Vanguard’s VCN as a replacement for XIU. VCN is more broadly diversified and has a lower MER.
Conclusion
It is difficult to replicate stock indexes at lower costs than the MERs of the best index ETFs. It would likely take a multi-million dollar portfolio and a complex strategy to get total portfolio costs below these MERs. And even then, a complex strategy would give only slight savings. Why bother? I prefer to just buy the best index ETFs.
One more cost to add to this calculation: Dan's time in learning all this and actually doing it. By the same reason I do not bake bread.
ReplyDelete@AnatoliN: Agreed. I'm doubtful that there is a way to save money even before valuing Dan's time.
DeleteLooking a it your way, VCN is quite a bargain for hoding 247 stocks at a 0.12% management fee. Great analysis!
ReplyDelete@Anonymous: VCN looks good to me, but I'm interested in hearing alternate points of view in case I've missed something. Thanks.
DeleteSome ETFs use sampling to replicate the performance with only a much smaller subset of stocks.
ReplyDelete@CanadianInvestor: True. But as I pointed out in the post, this reduction in diversification has a cost.
DeleteConcentrated asset portfolios exist and make sense. While sector diversification is obvious, sector asset representation need not mean owning all the stocks of that sector. XIU has six banks. Does that really add to diversification? Is it really necessary to own all the financials to have a balanced portfolio? It would say not.
Delete@Jon Evan: Concentrated portfolios only make sense for those who have an expectation of positive alpha after expenses. This can only be true of a very small minority of investors. Owning 6 banks is more diversified than owning one bank. Having six banks offers nothing in terms of sector risk, but it does reduce company risk. The vast majority of people who think they can pick the best bank are just overconfident.
DeleteIncreasingly, there are the DIY dividend investors (usually retired) seeking income and not alpha who use concentrated dividend stock portfolios. They buy and hold sector diversified assets and like the fact of no annual costs. In unregistered accounts, the 'clean' stock dividends make DIY bookkeeping/tax time easy unlike the 'dirty' XIU returns which complicate due to their ROC and other income returns. Regarding banks, they buy Canada's premier bank (TD), Canada's biggest bank, and the internationals BMO and BNS.
Delete@Jon Evan: I wish these people the best of luck. I'm sure that most don't understand the risks they are taking. Your emotional characterization of XIU returns is nonsense. It is far easier to do your taxes owning just XIU instead of owning a pile of stocks.
DeleteNicely explained. I have to say I have never understood the appeal of unbundling low-cost ETFs with many holdings like XIU. As you point out, we're probably talking about a couple of hundred dollars savings on $500K holding, and even then, that's only if you execute perfectly, which no one will actually do.
ReplyDeleteThe only time I could ever see this making any sense is with something like a Canadian REIT ETF, where you could probably replicate it very closely with 10-15 holdings and avoid a fairly high fee of 0.60%. But even then, it's a bit like doing your own oil changes to save $40.
@CCP: Thanks. I'm inclined to just wait and see if a competitor ETF comes along to reduce MERs. That' sort of what happened with VCN being an approximate replacement for XIU+XCS.
DeleteWhat are you thoughts on HXT versus XIU. I have HXT because it`s comm free through scotia.
ReplyDelete@Chris: I decided to wait and see about HXT and its swap structure and CRA's take on it. So far I've voted with my own dollars on VCN because of the greater diversification.
Delete