Last week I announced the 2012 returns for my model portfolios. Now it’s time to present the updated longer-term returns of the portfolios, with data provided by Justin Bender. This information covers the 15 years from 1998 through 2012:
Global Couch Potato (ETFs) | Complete Couch Potato | Über-Tuber | |
1 year | 8.07 | 8.71 | 9.01 |
3 years | 5.57 | 7.67 | 6.22 |
5 years | 2.62 | 4.76 | 4.00 |
10 years | 5.49 | 7.54 | 6.86 |
15 years | 4.77 | 6.66 | 6.06 |
Standard deviation | 7.76 | 7.76 | 7.93 |
Note these returns use actual ETF performance wherever possible. However, none of the funds has a 15-year track record (a few have been around for less than five years). So we have filled in the gaps using index data, minus the MER of the relevant ETF. It’s certainly not a perfect measure (no backtest is perfect), but it’s a reasonable proxy. Full details are provided in the PDF document, which is also linked on the model portfolios page.
A few observations about these numbers:
- As I’ve discussed before, start and end dates mean an awful lot. Unlike the returns for the decade ending in 2010, the 10-year returns now look quite good—so much for the “lost decade,” which only applies when your start date is during the tech wreck of 2000–01.
- Five-year market returns are disappointing, since they start just prior to the disaster of 2008. But barring a terrible 2013, the next batch of five-year returns will probably look outstanding, since they start with 2009, one of the all-time best years for equities.
- The Global Couch Potato lags the other two portfolios significantly over all periods longer than three years. This is largely because the last 15 years have seen strong returns in several asset classes that are absent in the Global Couch Potato: real-return bonds (9% annualized since 1998), Canadian REITs (13% since 1998), emerging markets (8.8% since 1999).
Dividend ETF webinar Thursday
On Thursday, January 17, at 11 a.m. EST, I’ll be participating in a webinar at ETF Insight. I’ll be joined by Vinit Srivastava of S&P Dow Jones Indices, Paul Kaplan of Morningstar CPMS, and moderator Yves Rebetez. The topic is Canadian dividend ETFs.
Specifically we’ll be looking at these ETFs tracking indexes from S&P Dow Jones and Morningstar:
iShares Dow Jones Canada Select Dividend (XDV)
iShares S&P/TSX Canadian Dividend Aristocrats (CDZ)
iShares S&P US Dividend Growers (CUD)
First Asset Morningstar Canada Dividend Target 30 (DXM)
First Asset Morningstar US Dividend Target 50 (UXM)
iShares Morningstar Dividend Yield Focus (XHD)
If you’re interested in learning more about dividend ETFs, I hope you’ll join us for the webinar.
Is this based on a 60/40 split? What would the returns be for other asset mixes?
@ Ryan: the model portfolio holdings and weightings are available under a tab at the top of the site.
It’s interesting to see that value and small cap tilt in the Uber-Tuber under performed the complete couch potato over the last 15 years…
Also, it’s good to see that the Complete couch potato slightly outperformed Mawer’s balanced mutual fund, which has a similar asset allocation and low management fee, over the last 15 years.
http://www.theglobeandmail.com/globe-investor/funds-and-etfs/funds/summary/?id=17966
@Jas: Yes, the small and value premiums are unreliable. Over the last 10 years, large has outperformed small in Canada, though not in the Us or internationally. Value outperformed growth in Canada and internationally, but not in the US.
@Jas – Yes and as the Mawer Balanced Fund is actively managed so it could only have a tax efficiency equal to or lower than the Complete Couch Potato. Of course not an issue in registered accounts.
@CCP – I didn’t see this noted on the pdf. Were these returns were calculated using annual rebalancing?
@Noel: Yes, the returns assume all portfolios are rebalanced annually.
Dan – did you calculated the long term returns for the “yield-hungry” portfolio?
@GeoEng: Unfortunately its impossible to get long-term index data for many of the asset classes in the Yield Hungry CP. Also, the nature of the distributions (return of capital, interest, dividends) is hugely important to the after-tax returns on that portfolio, and these are impossible to backtest and unpredictable in the future. I’ll have more to say about this in a post I’m preparing for next week.
@Jas – The Uber-Tuber’s 15-year underperformance relative to the Complete Couch Potato has nothing to do with the small and value tilts and everything to do with the exclusion of a large real return bond and Canadian REIT allocation (as well as a lower term to maturity on the bond allocation).
Is there a spreadsheet that we could have that would show each year’s return and allow us to change the weighting to use our specific examples or play around to see the differences?
@Justin: re: your dissection of the relative underperformance of Uber-Tuber due to the missing Real Return and Canadian REIT as well as the shorter term Bond allocation, if we exclude these drags from our calculations, how did the Fama French equities perform relative to the Plain Vanilla equities in Complete Couch Potato over the past 15 years? CCP tells us that the Canadian large-caps outperformed the small-caps over the past 10 years (but not the US and International large-caps) and he describes the small and value premiums as “unreliable”. But over the longer term would you agree with this characterization?
@Paul: No I don’t, sorry. The calculations were done with a database, not a spreadsheet.
@Oldie: By “unreliable” I simply meant that the premiums do not show up over all periods. This isn’t a controversial statement. There have been many periods of 10 or more years where growth has outperformed value or large has outperformed small.
@Oldie – I’ll create a more “apples-to-apples” comparison in the next couple days and post the results. I can tell you that I am about 99.99% sure adding a small cap and value tilt over the past 15 years would have resulted in a higher return, relative to a broad-market strategy (across Canadian, US, and International stocks).
I agree with Dan that the small cap and value premiums are unreliable…there is no guarantee over ANY period of time that these will be rewarded with higher returns over and above a broad-market strategy – there is also no guarantee that stocks will outperform bonds over any given period of time.
@Justin:
“I can tell you that I am about 99.99% sure adding a small cap and value tilt over the past 15 years would have resulted in a higher return, relative to a broad-market strategy (across Canadian, US, and International stocks).”
Yes, that is what I understood from a layman’s reading of the Fama-French literature, and the review of the UberTuber results over 10 years is not incompatible with my understanding of the academic predictions. I am in the dark about what is the actual value of the premium predicted by a Fama-French tilt when projected to infinity, because that would affect one’s calculation as to how much benefit one would try to capture by addition of such ETF’s, given that their MER’s are generally quite a bit higher. Also, I’m guessing here, but I would consider the volatility of a Fama-French type portfolio to be higher, or at least not lower than a corresponding plain Equities/Bonds Portfolio. Am I correct about this?
@Oldie – in the financial projections we are currently running for our PWL managed clients, we are including about a 1% small-cap/value premium above the market. In other words, if you are expecting broad-market equities to return 7% over the long term, our projections would assume about 8% for a small-cap/value strategy.
I look at this strategy from a different risk reduction point of view than most advisors; let’s assume that you have gone through the planning process and discover that you require a 5% rate of return to meet your goals. Further assume the following expected asset class returns:
Bonds = 2% expected return
Broad-market stocks = 7% expected return
Small-cap/value tilted stocks = 8% expected return
To have the greatest chance of meeting your goals, you would need to take on 60% broad-market stock risk and 40% bond risk (7% x 0.6 + 2% x 0.4 = 5%). If you instead chose to tilt toward small-cap and value companies within your stock allocation, you could reduce your equity allocation to 50%, and increase your bond allocation to 50% (8% x 0.5 + 2% x 0.5 = 5%), thereby reducing the volatility of your portfolio without decreasing the expected return.
@Justin Bender: I get it. And, implicit in your strategy, I guess, is the underlying principle that the volatility of a portfolio with 50% small-cap/value tilt equity portfolio is no higher than a portfolio with 50% broad-market stocks?
I understand your estimates are just that — estimates, but are these estimates of expected returns before MER costs? Because, just eye-balling the MER’s of the small-cap/value ETF’s that have been discussed for consideration, they all seem to be in the range of 0.3 to 0.5% higher than the broad market stocks available, and that differential will eat into your profits.
@Oldie – A 50% small-cap/value tilt equity portfolio would be expected to have higher volatility than a portfolio with 50% broad-market stocks (but lower than a portfolio with 60% broad-market stocks).
The estimates are before MERs or taxes. I would definitely agree that an investor should carefully consider all the additional costs of an alternative investment strategy.
For DIY investors, I’m a big advocate of keeping things simple; low-cost, tax-efficient, broad-market index ETFs and mutual funds are an excellent choice (once the investor has determined an appropriate asset allocation, based on their ability, willingness, and need to take risk).
@Justin
Where does one find useful information about “expected returns” of the various asset classes?
@DaveL – I normally use the Shiller CAPE ratios for Canada, US, International, and Emerging Markets. To estimate the real expected return, I take the inverse of the Shiller CAPE, and adjust it downwards by about 1% (to account for future earnings “slippage”). Slippage is the result of net issuance of stock over time, resulting in less earnings for existing shareholders. You would then have to add your inflation expectation in order to obtain a gross return.
Our Director of Research, Raymond Kerzerho, provides me with these Shiller CAPEs on a quarterly basis. The only website I’ve seen similar numbers published in one place occasionally is on Mebane Faber’s blog, “World Beta”: http://www.mebanefaber.com/2012/10/24/sector-cape/
For government bond ETFs, the best estimate to use would be the yield to maturity of the product minus the MER.
Of course these are just estimates (guaranteed to be 100% inaccurate), but investors would be wise to pay attention to the differences in expected returns between stocks and bonds over time. For example, if an investor requires a 4% return to meet their goals, and the expected return on both stocks and bonds is 4%, they may want to consider a more conservative asset allocation, as they are not expected to benefit from increasing the allocation to stocks.
@Justin Bender: just to clarify,
“Small-cap/value tilted stocks = 8% expected return”
“Small-cap/value tilted stocks” would be, for example as in the UberTuber Portfolio, where VTI and VBR are held in 2/3-1/3 ratio for the US component of equity, right? I mean you don’t have to hold zero VTI and all VBR as the US equity component, and similarly in Canadian and International Equity to achieve this expected return, do you?
@Oldie – the expected return of 8% I quoted is based on the assumption of a 2% small cap premium and 2% value premium going forward and approximate small cap (SMB) and value (HML) coefficients (using a 3-factor regression analysis) of about 0.25 for each factor.
Example:
Expected Premium Above Market:
= Small Cap Premium x SMB coefficient + Value Premium x HML coefficient
= 2% x 0.25 + 2% x 0.25
=1%
I would expect the Uber-Tuber portfolio to have around the same expected return as the portfolios I construct for clients (based on the same assumptions I used).
Tiss true, timing is everthing. To make real money from the indices above inflation, an investor would have to have been part of the raging bull of the 80’s and 90’s, as I outline in my article “No one makes money from the stock markets”.
http://seekingalpha.com/article/949001-almost-no-one-makes-money-from-the-stock-market-alone
RE: Dividend ETFs
Where do dividend ETFs fit in a Couch Potato Portfolio, if at all? Could a dividend ETF in conjunction with a small-cap/value ETF replace a Canadian index?
@Justin Bender: I hope I’m not making a nuisance of myself hogging the comments, but, still trying to refine my understanding of how potent the effect is of the small caps component, and how big an allocation of small caps is required to have that effect, given what a small percent of selected small caps equity (1/3 of each equity component) there is in, say, the Uber Tuber portfolio, what is the relative performance expectation of an ETF like ZCN which tracks the TSE 250+, versus, say, XIU, that selectively tracks only the top 60 stocks of the TSE. Given the relatively lower weighting of the sum of the smaller caps (that is, those smaller than the top 60) represented in TSE250+, is this lower contribution so small as to make hardly any difference to the performance?
@Tasty Crumb: I think it’s reasonable enough to use dividend ETFs in lace of broad-market funds as long as you pay attention to sector diversification (dividend funds tend to be overweight some sectors and underweight others relative to the overall market). But it’s not something I typically recommend, especially in a tax-sheltered account.
@Justin Bender: perhaps that was a poor example, given the less than ideal characteristics of the Canadian small-caps in general; suppose there was an equally priced ETF to VTI (which tracks the NYSE top 3,300) that tracks the S&P 500; how would you assess their relative expected risks and returns?
@Oldie: Justin may want to add something, but I can tell you that the relative performance of the large-cap S&P/TSX 60 and the broad-based S&P/TSX Composite has been small but significant over the long term.
Interestingly, over 25 years, the S&P/TSX 60 has outperformed (8.99% versus 8.27%). The indexes are almost identical over 10 years (9.28% versus 9.21%), and over three and five years, the Composite has outperformed be about 1% and 0.5% annualized, respectively.
The small-cap outperformance is greater in the U.S. (based on the Russell 1000 and Russell 2000 indexes) over the medium term: about 2.2% over 10 years and 1.1% over 15 years. But going out 20 and 25 years the performance of the two indexes is extremely close.
@Oldie – I ran a quick-and-dirty monthly regression analysis from January 2001 to December 2012 for a number of Canadian indices – here are the small cap coefficients – as would be expected, the small cap regressions increase as you move from larger cap indices to smaller cap ones:
S&P/TSX 60 = -0.16
S&P/TSX Composite = -0.02
S&P/TSX Completion = 0.33
S&P/TSX Small Cap = 0.85
For more information on the process involved, feel free to visit William Bernstein’s site: http://efficientfrontier.com/ef/101/roll101.htm
@Oldie – I’d most likely compare IVV to VTI (since June 2001), compile the monthly returns in USD, download the 3-factor regression factors from Ken French’s website, and run a regression to determine the small cap and value coefficients for each fund. VTI would most likely have small cap and value coefficients equal to 0 (since it is basically “the market”), and IVV would most likely have a negative small cap coefficient (perhaps -0.15?) and a value coefficient equal to 0. If I assume 2% small cap premium going forward, the expected return of IVV would lag VTI by about 0.30% per year (-0.15 x 2%).
@CCP,@Justin Bender: Thanks guys, this is very illuminating, and illustrates the subtle variances from theory to reality, or, in the Canadian situation, the need to factor in all the peculiarities of the particular situation to understand how theory can sometimes break down.
@Justin Bender: If I understand the calculations correctly, the 2% premium of the small-cap population is relative to the general broad market population. And, several comments back, the Value stock population likewise is associated with a 2% premium relative to the broad market population.
If this is so, do I also understand correctly that the reasoning behind combining the two types to fill up an Equity portion (for instance in the UberTuber Portfolio Canadian equity which has 12% iShares Canadian Fundamental CRQ and 6% iShares S&P/TSX SmallCap XCS) is to exploit this 2% premium and somehow manage the risk better than having only one component. I suspect that this pairing somehow also exploits some characteristic of the higher individual risk associated with each of these asset sub-classes. Do their variabilities have different correlations from each other, or have some similar cross-effect to lower the combined risk?
@Oldie – the 2% small cap and value premium is relative to large companies and growth companies respectively (not the broad-market). The 2% premium is the expectation I currently use…I’m sure if you talk to 10 different advisors, you’d get 10 different expectations. Keep in mind that the actual premiums will certainly be different than 2% over your investing time horizon…they could even be negative. In other words, tilting towards small cap and value companies does not guarantee a higher return.
As Dan mentioned, the small cap and value premiums are unpredictable and can show up at different times. In 2012, Canadian value stocks outperformed Canadian growth stocks by a significant amount, but Canadian small cap stocks underperformed Canadian large-cap stocks by a significant amount.
@Justin Bender: Yes, I do understand that the theory is only a prediction, and that, depending on the standard deviation, you can experience a lot of negative and positive swings before tending to the predicted average gain towards infinity time if the theory is correct and is still accurate; or not at all if it isn’t!
Thanks for keeping us updated Dan. The more basic math that people can see with their own eyes, the more converts common sense investing will win. I actually see advertisement on American networks these days that seek to educate retail investors about how ridiculous management fees are, and I think that is a huge step. Now if only we could get more Canadians to see the light. The most popular articles on my site continue to be people and advisers defending the mutual fund industry in the comment section. It’s unbelievable how thorough their propaganda is. Thanks again for keeping the basic math front and center in this argument. “Lost decade” for who right? I’ll take that 15 year average all day long, especially when you consider a “great recession” is bookended in there.
Hi CCP, I realize that backdating results for the Yield Hungry Couch Potato would be difficult (if not impossible), but do you have results for the time period that you have published it? (or even just the last couple of years?)
@Greg: I calculated the 2011 and 2012 returns only:
https://canadiancouchpotato.com/2012/01/09/couch-potato-portfolio-2011-returns/
https://canadiancouchpotato.com/2013/01/11/couch-potato-portfolio-returns-for-2012/
The individual fund returns are available on their respective web pages, and you might be able to go back another year or two.
I currently have a RSP with ING – Streetwise Balanced Portfolio and a high interest TFSA savings account with ING.
I have opened a TD-eseries TFSA account too.
Should I convert the high interest TFSA account at ING to a Streetwise account and at the same time have a TD-eseries TFSA account?
Does it make sense to have two TSFA mutual fund account.
1. streetwise balanaced
2. td – eseries
Both will get equal contributions.
@Rocky: I don’t see any benefit in having a TFSA with two different companies. It’s usually best to simplify your life by just using a single account.
ok.
one more question, as my RSP is in ING Streetwise.
would it be better to move my TFSA mutual funds to TD e-series? Won’t having TFSA in ING Streetwise be like having all my eggs in one basket?
Also in future, If I max out my TFSA mutual fund limit (assuming it’s in ING), should I create a mutual fund account with the same bank (ING) or pick another bank (TD-eseries)?
Thanks a lot of your help.
@Rocky: You don’t need to worry about having all your accounts with one financial institution. All of these firms are member of the Canadian Investor Protection Fund (CIPF), which protects you in the case of insolvency.
My concern was more with the performance.
i.e. if the ING portfolio goes down then TD e-series might go up and things get balanced.
Is that a good idea?
@Rocky: If you are planning on using the Global Couch Potato with e-Series funds, the performance will be very similar, since the GCP and the ING Streetwise Balanced Portfolio have almost the same holdings.
The best advice I can give you is not to overthink this. Pick one institution and one simple portfolio and concentrate on saving.