Canadian Money Forum banner

281 - 300 of 304 Posts

·
Registered
Joined
·
63 Posts
I haven't gone through all 14 pages, but just wondering if anyone has backtested this approach for the previous 10 years/25 years and compared the returns to the benchmark index? I see XIU has returned 6.5% over 10 years and 6.61% since inception.

I'm currently invested mainly in VGRO for the simplicity, but would love to add a small portion of my investments to another theory similar to this for a chance at slightly higher returns.

That being said.. is there even any easy way to back test theories like this or similar ones? I've always just trusted others that have said they've backtested, such as Canadian Couch Potato portfolios, but it would be interesting to be able to backtest my own portfolio ideas.

Edit: Also just got to the post where you mention this is all in non-registered accounts for tax purposes and you hold index funds in TFSA/RRSP. I'm not at a point where my TFSA is maxed yet so I guess I will be better off continuing to add to VGRO until both are maxed?
 

·
Registered
Joined
·
195 Posts
I haven't gone through all 14 pages, but just wondering if anyone has backtested this approach for the previous 10 years/25 years and compared the returns to the benchmark index? I see XIU has returned 6.5% over 10 years and 6.61% since inception.

...

That being said.. is there even any easy way to back test theories like this or similar ones? I've always just trusted others that have said they've backtested, such as Canadian Couch Potato portfolios, but it would be interesting to be able to backtest my own portfolio ideas.

To backtest, use Portfolio Visualizer

The original posts talks about a 5-pack consisting of RY, FTS, CNR, ENB & BCE. And, yes, it definitely outperformed XIU. But I'll let James talk about his own experience.

What if the selection was different, but in the same industry? What about TD, EMA, CP, TRP & T? About the same results.

Backtesting doesn't guarantee future result, but I agree that it's good to see historical trends and past performance.

I guess selecting these as a 10-pack would've been more prudent.

20377


20378
 

·
Registered
Joined
·
17,004 Posts
I haven't gone through all 14 pages, but just wondering if anyone has backtested this approach for the previous 10 years/25 years and compared the returns to the benchmark index? I see XIU has returned 6.5% over 10 years and 6.61% since inception.
If you mean my method, yes, I have back-tested it over about 17 years + 3 years forward testing in real life and it has outperformed XIU. But my original goal was not to outperform XIU, it was to roughly replicate XIU performance.

I still don't think my 5-pack will keep outperforming XIU. Instead, I see it as a replacement for it, using individual stocks instead of the ETF.

That being said.. is there even any easy way to back test theories like this or similar ones? I've always just trusted others that have said they've backtested, such as Canadian Couch Potato portfolios, but it would be interesting to be able to backtest my own portfolio ideas.
It isn't easy to back test. It's hard because you have to make decisions the same way you would have made them at that previous point in history, and when doing back tests we inevitably carry some hindsight knowledge and bias. In my case for example, the sector selection is tainted by hindsight bias.

Similarly with the Canadian Couch Potato, he's used certain allocation to countries but you'll notice he doesn't have a massive emerging markets weight. That is also hindsight bias. It's possible that if they started that method in 2003, he might have used a big emerging markets weight.

If the Canadian Couch Potato started doing this in 1989, he might have had a big Japan weight. But when we do this after the fact, everyone pretends they would have avoided Japan. Nope.... that's bias. Everyone in the late 80s invested heavily in Japan. Just like everyone today invests heavily in the USA.

Back tests have their limitations. There's always some hindsight bias.
 

·
Registered
Joined
·
17,004 Posts
To backtest, use Portfolio Visualizer

The original posts talks about a 5-pack consisting of RY, FTS, CNR, ENB & BCE. And, yes, it definitely outperformed XIU. But I'll let James talk about his own experience.
That still is not a correct back test because those particular stocks were not the stock picks ever since the start date. The portfolio changes over time, so you also have to back test the management process (which stock is held when). It's quite complicated.

Back testing index portfolios is much easier than back testing individual stock picks. With indexes, you just look at trailing performance.

With individual stocks you have to carry out the portfolio management operations, the buys & sells.
 

·
Registered
Joined
·
195 Posts
I still don't think my 5-pack will keep outperforming XIU.
I think you'll greatly outperform XIU.

XIU is being dragged down by stocks that became bad performers. At least SHOP helped a bit, but SHOP won't always be there.

While your selection always had a steady increase or a great dividend.

Meanwhile, ABX is still high in the XIU holdings because of its market cap. It has helped only during the crash. An index distribution only based on market cap is definitely oversimplifying.

That still is not a correct back test because those particular stocks were not the stock picks ever since the start date. The portfolio changes over time, so you also have to back test the management process (which stock is held when). It's quite complicated.

Back testing index portfolios is much easier than back testing individual stock picks. With indexes, you just look at trailing performance.

With individual stocks you have to carry out the portfolio management operations, the buys & sells.
I though you kept the same 5 stocks and same equal weight rebalanced every year? Sorry.

You can input a cashflow in PortfolioVisualizer, but since your portfolio and XIU are both low volatility, historical the 5-pack easily outperformed XIU.
 

·
Registered
Joined
·
17,004 Posts
While your selection always had a steady increase or a great dividend.
Dividends are not a factor in total returns. They are totally meaningless from a performance standpoint. Each dividend knocks down the share price.

See this video from Ben Felix: The Irrelevance of Dividends
And this one that describes that Dividends do not matter

I though you kept the same 5 stocks and same equal weight rebalanced every year? Sorry.
Nope, it's impossible to pick a constant group of stocks and sit on them for 20 years. You have to do some kind of portfolio management, and my management procedure is described in post # 190. You can even see a coloured diagram of the change in holdings over the years.
 

·
Registered
Joined
·
195 Posts
Dividends are not a factor in total returns. They are totally meaningless from a performance standpoint. Each dividend knocks down the share price
I mean that the total return was a steady climb. Either it's a no dividend stock growth 5% or a 5% dividend stock with no growth. We've had this debate somewhere else. Post #133 and Post #144

In the first minute of your Ben Felix video, he points out "I do not mean that they are not an important component of total returns". I'm talking about total return stability. He's just saying that a good dividend doesn't mean a good stock, which is obviously true.

Nope, it's impossible to pick a constant group of stocks and sit on them for 20 years. You have to do some kind of portfolio management, and my management procedure is described in post # 190. You can even see a coloured diagram of the change in holdings over the years.
Well, it is possible, but most likely improbable since after many years you maybe find other stocks that - you think - are better options and should replace your current stock.

From your diagram, three of the 5-pack have been kept most of the time. Therefore, it has been possible for at least 3 of your 5-pack, that's just what I meant. So it could've been possible for 5 out of 5.

Selecting SU in 2009 when it crashed was the rational decision after its huge run, compared to EMA, but you could've also decide to stick with EMA with its 9% CAGR from 2001 to 2009 when seeing that SU had a few crashes in its history (1997, 2000, 2002) before its huge run. But since SU totally outperformed EMA from 2002 to 2009, well I totally understand the idea to switch to SU.

I was just trying to say that would could backtest and pick stocks that had a 20-year history of steady growth and stick to it unless there's some really bad event. As of now, FTS had a steady growth for over 20 years, big 6 banks had a steady growth for over 20 years, CNR had a steady growth for over 20 years so these are not going nowhere else unless something huge happens.

By the way, I think your 5-pack is awesome, I'm not challenging your decisions and I'm nowhere near a good position to challenge them, I just wanted to point out that how your experienced your 5-pack is one scenario out of many other possibilities. Some other scenarios would've been worst, some would've been better, but you never know and it depends on your strategy, your beliefs, your context and information at that moment, your personality and much more.

It is just not "impossible" that one stick to the same 5-pack during 20 years. Maybe not likely, but not impossible. Nothing would've stop someone to pick a 5-pack and keep being satisfied with its selection during 20 years, Did you do that? No. Will I do that? No. Can someone out there do that? Yes.
 

·
Registered
Joined
·
195 Posts
We may say that making a choice that will keep us happy during 20 years is a bit lucky but...

Not to make a bad joke, but after all, there's people out there still with their spouse after 50 years...

Ok, that's a bad analogy.
 

·
Registered
Joined
·
392 Posts
@james4beach I wonder how you have done "proper" backtesting?
I mean where you actually program the algorithm/ruleset of stocks selection.

It looks like Quantopian now supports Canadian equities data, so I might give it a shot later to backtest your formula.
If I remember correctly it was simply: "Pick the highest market capitalization stock from each of the 5 sectors at the beginning of year/rebalancing time" ?
 

·
Registered
Joined
·
3,865 Posts
Why not go back to the criteria Argonaut originally posted to achieve a 6 pack that easily beats the performance of the TSX. This 5 pack thing of James is a more recent copy with different criteria. I'd link the original post but am too lazy...its there if you search though.
 

·
Registered
Joined
·
195 Posts
I tried to find the stocks with the most stable growth during 20 years.

Couldn't find 5 or more. Ended up selecting CNR, FTS, CAR-UN and MRU. Waiting to find more.

CNR and FTS are definitely the winners with CNR having the best growth and FTS having the least drawdown. MRU having the worst drawdown. CAR-UN having the longest drawdown.
 

·
Registered
Joined
·
17,004 Posts
@james4beach I wonder how you have done "proper" backtesting?
I mean where you actually program the algorithm/ruleset of stocks selection.
My algorithm can be found in post # 190. It's technical and based on looking at XIU (or the TSX 60), not based on earnings or fundamentals.

I did the backtesting by pulling up 17 years of financial statements for iShares XIU and applying my stock selection criteria to each Dec 31 snapshot of the portfolio. I pretended that I saw the XIU holdings on Dec 31 and made the portfolio changes in response.

I'm not familiar with the software you mentioned, but maybe it can do it.
 

·
Registered
Joined
·
17,004 Posts
I thought you might have used Quantopian: The Place For Learning Quant Finance
It seems to have only recently acquired non-us (including Canadian) stocks data.
It's a lot of fun to play with, if you have basic python skills and think you can "beat the market" :)
Neat, I have not ever played with those. I keep a proprietary stock database of my own, with data that I slurp from various places for about 800 stocks and ETFs.

Beating the market? Sure... give me a hundred tunable parameters and I will build you the most successful stock prediction system ever made. It will be totally useless garbage, though ;)

(for those curious about what I mean, see this article on over-fitting)
 

·
Registered
Joined
·
400 Posts
The 5-pack (Ry/FTS/CNR/ENB/T) vs ZLB look to have the same performance over time.... as far as I can see unless I am missing something....

Simplicity says just buy ZLB no ?
 

·
Registered
Joined
·
2,975 Posts
The 5-pack (Ry/FTS/CNR/ENB/T) vs ZLB look to have the same performance over time.... as far as I can see unless I am missing something....

Simplicity says just buy ZLB no ?
But there's an MER cost, along with tax distributions of ROC, Other Income, Foreign Income, Capital Gain along with about half that's eligible dividends is there not?

Wouldn't it be easier to just own the stock at no cost and a simple T5 slip?

ltr
 

·
Registered
Joined
·
195 Posts
The 5-pack (Ry/FTS/CNR/ENB/T) vs ZLB look to have the same performance over time.... as far as I can see unless I am missing something....

Simplicity says just buy ZLB no ?
Well, I'd say this thread is about the 5-pack approach which mostly points out how managing only 5 stocks can outperform the index.

But if you are looking for performance through ETFs, well that's another topic and there's many options, many even better than ZLB, even though ZLB is a very good choice.

But there's an MER cost, along with tax distributions of ROC, Other Income, Foreign Income, Capital Gain along with about half that's eligible dividends is there not?

Wouldn't it be easier to just own the stock at no cost and a simple T5 slip?

ltr
But ZLB is all Canadian holdings, I guess it simplifies? I'm a beginner stock investor and it's currently all in TFSA, so I still don't know in what kind of tax mess I'll get into when I'll buy international ETFs for one of my portfolios...

[EDIT] Nevermind, I just downloaded ZLB's distribution tax and yes there's all that. How come since it's all Canadian holdings? No? Ah, is that because these companies make money internationally?
 

·
Registered
Joined
·
17,004 Posts
The 5-pack (Ry/FTS/CNR/ENB/T) vs ZLB look to have the same performance over time.... as far as I can see unless I am missing something....

Simplicity says just buy ZLB no ?
You're right, they are similar. I think this is because both exclude commodities and resources so both are essentially holding the TSX, underweight commodity stocks. In other words, they have similar sector exposures. Both are very light on tech.

I have considered buying ZLB instead of my individual 5 stocks and I think one could make that substitution. ZLB is better diversified, which is a plus.

In my non registered account, I prefer to hold these individual stocks. I prefer the tax accounting with the individual stocks, since I don't have to deal with the ETF distribution mess. I really prefer keeping Canadian ETFs inside registered accounts only.

But there's also a hefty 0.39% fee on ZLB. For my 100K invested, assuming I make just 5 rebalancing trades a year @ $10 each, my "MER" works out to just 0.05% which is a fraction of the cost of ZLB.
 

·
Registered
Joined
·
400 Posts
So, use ZLB in RRSP/LIRA/TFSA/ect and in the MARGIN/CASH acct use the 5-Pack ? Seems to make sense to me... simplicity has to be worth something...

J4B have you done a backtest on your 5-Pack vs ZLB ?
 
281 - 300 of 304 Posts
Top