Canadian Money Forum banner

1 - 20 of 41 Posts

·
Registered
Joined
·
1,516 Posts
Discussion Starter #1 (Edited)
I've lately been reading a great deal about reducing risk by building a portfolio of non-correlating investments. In fact, it's been demonstrated that you can often lower the standard deviation by combining a risky but non-correlating investment to your asset mix. The trouble seems to be that the world markets are becoming more and more correlated.

I guess the most obvious non-correlating assets to stocks are fixed income products. These deserve a place in the portfolio, but the return is currently fairly low. A book I'm currently reading recommends hedge funds for a portion of the portfolio but these products seem generally more suited to the more wealthy and as well, can be extremely dangerous if chosen poorly. Real estate may be another possible choice, but as we've seen in the last meltdown, REITs were definitely not exempt from the past carnage. However, in many parts of the country, residential real estate held up quite well. I've never been a gold bug, but am currently considering whether it needs a small place in my portfolio -- however, as we saw, it was also not exempt from the last crash.

So here is my question. What products do you use for negative correlation to Canadian stocks? Are there any mutual funds that you are aware of that either tend to have a rather negative correlation to the market or which were relatively unscathed by the last crash? Any stocks that you find generally run counter to the market? Any other products that you use for this purpose?
 

·
Registered
Joined
·
3,197 Posts
Non-correlating" sounds like a technical descriptor for why you should diversify by asset class, geography, sector, etc. I hadn't though of Income vs. Equity as an example of "non-correlating', but I guess it is.

In a sense, any mutual fund that invests across all sectors is already employing a technique of mixing non-correlating assets.

You make an interesting observation about world markets becoming more correlated, which brings into the question the conventional wisdom about the importance of diversifying your equity by CDN/US/International. Consequently it might make more sense to focus on diversification by Large Cap/Small Cap and Growth/Value than by geographic region. From a CDN perspective however, there aren't a lot of good choices in CDN Small Cap equity, because of our small market size. And the distinction between growth & value stocks is often highly subjective - the terms tend to be used as marketing adjectives in a prospectus rather then anything measurable.
 

·
Premium Member
Joined
·
2,686 Posts
This is a great topic guys!

Here's some correlation data (a bit dated though) from Bylo's website:

http://www.bylo.org/cancorel.html

While it is true that global markets are becoming more correlated, there is still diversification benefits to be had from adding foreign stocks to a portfolio. The same goes for REITs. Yes, REITs fell along with equities in the 2008-09 bear market but I believe REITs and stocks moved in opposite directions in the 2002-03 bear market.

Fixed-income and stocks have low correlation and though fixed income has low expected returns, you can make up some of the shortfall through diligent rebalancing. The reason to add fixed income is to lower volatility and a small reward penalty might be worth it for most investors.
 

·
Banned
Joined
·
419 Posts
Most of my dividend growth stocks run contrary to the market.
People get spooked and flock to DG stocks and drive the price up.
Very. Expensive right now.
Guess I'll have to wait until the next craze, then my DG stocks won't be "cool" anymore.
 

·
Registered
Joined
·
2,626 Posts
This is a great topic guys!
One of your favorite ones right CC? ;)

One thing to keep in mind is that (as highlighted by the link) correlation among asset classes shifts over time - so I think it's important to constantly evaluate your allocations.

Corporate bonds and stocks have had a high correlation in direction of movement over the past 2 years.
 

·
Registered
Joined
·
3,140 Posts
The problem you have with adjusting your portfolio so that it is "uncorrelated" is three fold.

1) Most asset classes move in the same direction on any given day. By that I mean if you have 6 asset classes, 5 will probably go up and down in the same direction on any given day. For example Cdn. stock versus US stocks. The non-correlation comes into effect over 1 year or 3 years, etc. In other words, Cdn. stocks make 10% when US stocks only make -5%. This phenomenum makes the benefits of this procedure a little annoying, because we tend to look at our portfolio's more often then once every 1 year or 3 years.

2) If you find 6 asset classes that have reduced correlation, you will inevitably find one that, at the time, you will not like. For example, as you have said, "bonds". Hence, we tend to build our portfolios with everything that is working well at the time or that we think will do well, and of course, those things will be more correlated, for them to be working at the same time.

3) Lastly, even if you find an asset that is completely uncorrelated with another, it does not mean that if the one goes down 1%, the other goes up 1% or more. Usually what happens is one goes down 1% and the other goes up 0.05%. They are completely uncorrelated, but not very much help. Bonds aligned with stocks, would again be an example of this. In 2008, bonds went up, but not nearly enough to make them worthwhile considering how much was lost in stocks. Their benefits came by diluting the losses on your stocks, not by its inverse move.

Anyway, that is how I have seen things. Good luck to you.
 

·
Registered
Joined
·
249 Posts
Canadian Long Term Bonds (XBB.TO) had no correlation with the stock market in the last few years.
 

·
Registered
Joined
·
222 Posts
Given a set of "n" stocks that you are considering for your portfolio, the basic modern portfolio theory process is to:

- derive an expected return for each stock
- compute the standard deviation for each
- compute the correlation & covariance matrices based on the stocks' historical returns
- using these data, compute the n-way efficient frontier: this involves, in effect, solving a reasonably straightforward linear programming problem to determine the optimal set of weights for each stock in your portfolio.

I've done this by hand for 2 stocks. It's tedious/cumbersome, and unless you're particularly masochistic (or maybe you've been Focused by a group of Emergents and forced to solve accounting problems), I wouldn't suggest doing it by hand for n > 2. Thankfully, there are a number of freely-downloadable excel spreadsheets that will do this for you. I might suggesting googling for "markowitz frontier excel". I think that one example is here.

(That said, my general advice is that I don't think that you need to be quite this extravagant... a risk-optimal portfolio does not guarantee the best performance. While I'm fully aware of this technique, I tend not to use it myself and prefer instead to "eyeball" things by asset class to achieve my desired diversification.)

regards,
K.
 

·
Registered
Joined
·
2,925 Posts
Right now it is simple and that is US dollar, not stocks against everything else. I recently sold some gold for US dollars to hedge against a drop in the Canadian dollar and gold. When everything drops people have only one place to go and that is US dollars and maybe the Japanese Yen.
 

·
Registered
Joined
·
1,516 Posts
Discussion Starter #10
I'll just pass this on for anyone interested (another investor passed it onto me on another forum). Ishares has a very cool correlation calculator that will show the correlations between all your stock and mutual fund holdings and also show Ishares products that have negative correlations (or positive) with your portfolio:

http://ca.ishares.com/tools_charts/inde ... calculator

You have to register to access the software. For whatever reason, it did not allow me to access the software the first day I registered.
 

·
Banned
Joined
·
407 Posts
My 2 cents

The search for poorly or negatively correlated assets can be futile for reasons already stated. Correlations can change over time, so if you carefully construct your portfolio to be weakly correlated you can still be disappointed. Also, as previously stated, negative correlation is meant to reduce volatility, not increase return (it might increase your risk-adjusted return).

Some have stated that bonds are not correlated to stocks. That's not my understanding. Gov't bonds have a weak positive correlation to stocks while corporate (esp. junk bonds) have a strong positive correlation.

Someone mentioned hedge funds previously as well as the main drawback: you usually need a lot of money to invest. I'm sceptical; many hedge funds are glorified mutual funds with outrageous fees. The one exception to this may be hedge funds that are dedicated "short" funds. That is, they make their money by shorting stocks. These funds get killed in bull markets but make money in bear markets i.e. negatively correlated.

No one has mentioned inverse ETFs. These products are designed (generally using futures) to give inverse performances compared to an index. For example, Horizons Betapro has a number including ones that move inversely compared to the S&P 60, S&P 500, CDN financials, US dollar etc. However, these ETFs are generally used as trading vehicles, not as buy-and-hold assets. There are even ETFs that give you double inverse movement!

Theoretically, you could short futures on stock indices to match your long position, but that's getting complicated (and expensive) for most people.
 

·
Registered
Joined
·
2,054 Posts
I don't really think that the last crash is a great example to use when discussing correlation of assets because some things correlate to everything sometimes in unexpected ways. When the whole banking system collapses it affects everything.

When I think of correlation as it applies to stocks in general even in the same asset class. So for instance if you wanted to buy retailers you might want to buy both Dollarama and Tiffany & Co

If you buy miners then maybe Peregrine Diamonds and North American Palladium which mine entirely different metals.

You might want to avoid buying Barrick Gold and New Gold and Franco Nevada mines because they are all dependant on gold price.

This way no matter how the market turns you're still covered. One will go up the other goes down.
 

·
Registered
Joined
·
1,516 Posts
Discussion Starter #13 (Edited)
The topic came up after reading "How Harvard and Yale Beat the Market" by Matthew Tuttle. The author describes how endowments consistently beat the markets with lower volatility by combining investments with low correlations (perhaps I should have used the word "low correlation" rather than "non-correlation".) However, the goal of these endowments tends not to be to beat the market but rather to not lose money and to produce a cash flow. The author explains that the main point of combining assets with low correlation is to reduce or eliminate losses because losses tend to have a much more damaging effect on portfolios than gains have a positive one. Another author, William Bernstein demonstrates how you can combine two risky assets (in this case the S & P index and the international index) and, over the long run, end up with both higher returns than either index individually and a lower standard deviation (risk factor).

I guess when you whittle it all down we're talking plain old diversification

I found the calculator on the Ishares site informative. It shows, for example, that gold stocks, Walmart, pharmaceuticals and health care stocks tend to have either no correlation or a slightly negative correlation to much of the market in general and many REITs have a low correlation.
 

·
Banned
Joined
·
419 Posts
David Dreman points out in his book "Contrarian Investment strategies":

Rule #3:

Do not make investment decisions based on correlations. All correlations in the market, whether real or illusory, will shift and soon disappear.
 

·
Banned
Joined
·
407 Posts
David Dreman points out in his book "Contrarian Investment strategies":

Rule #3:

Do not make investment decisions based on correlations. All correlations in the market, whether real or illusory, will shift and soon disappear.
Great point! Dreman's book is an underrated, under-read classic.
 

·
Registered
Joined
·
1,516 Posts
Discussion Starter #16 (Edited)
David Dreman points out in his book "Contrarian Investment strategies":

Rule #3:

Do not make investment decisions based on correlations. All correlations in the market, whether real or illusory, will shift and soon disappear.

I will order that book from the library, as I always like to keep an open mind to all opinions. However, while I agree that correlations will often shift, I don't agree with the enitirety of that statement. To take the most obvious example, stocks will almost always not be well correlated with cash or short term bonds. Stocks in general will also usually have very low correlation to gold. As well, if one was to avoid investing because the fundamentals of those investments would eventually shift or disappear then one could use the same logic to not even invest in the stock market. After all, the majority of companies from the original DOW are now no longer in existence.
 

·
Registered
Joined
·
126 Posts
David Dreman points out in his book "Contrarian Investment strategies":

Rule #3:

Do not make investment decisions based on correlations. All correlations in the market, whether real or illusory, will shift and soon disappear.
Agreed. Buy undervalued companies with a margin of safety thats within your circle of competance is more than sufficient.
 

·
Banned
Joined
·
407 Posts
I will order that book from the library, as I always like to keep an open mind to all opinions. However, while I agree that correlations will often shift, I don't agree with the enitirety of that statement. To take the most obvious example, stocks will almost always not be well correlated with cash or short term bonds. Stocks in general will also usually have very low correlation to gold. As well, if one was to avoid investing because the fundamentals of those investments would eventually shift or disappear then one could use the same logic to not even invest in the stock market. After all, the majority of companies from the original DOW are now no longer in existence.
Actually most of the original DOW companies still exist - though not as DOW components or publicly traded companies (many were bought out or taken private). The question is, are correlations stable enough to make long-term assumptions? The answer is usually "no." And whether stocks correlate well with gold depends on the stocks themselves. Would you argue that gold stocks correlate poorly with gold? Canadian stocks, because of the high number of resource companies, actually correlate highly with commodities and gold while American stock indices do not.

I don't think any one said don't invest because correlations shift. I certainly didn't. What I am saying is don't rely on a false sense of security with "negatively correlated" assets.

Also, stocks and short-term junk bonds correlate highly.
 

·
Premium Member
Joined
·
2,686 Posts
Agreed. Buy undervalued companies with a margin of safety thats within your circle of competance is more than sufficient.
This is easier said than done. Buffett makes it appear that stock picking is an oh-so-easy endeavor. It's not. Buffett is a genius with an encyclopedic knowledge of companies and industries and an incredible calculator of investment odds. Perhaps 1% of professionals can match Buffett's skill. Joe Public doesn't stand a chance. Doesn't mean that there are no gifted amateurs. Just that you should make sure you fall in the gifted category.
 
1 - 20 of 41 Posts
Top