Asset Allocation - YouTube

There are few things that we can  control when it comes to investing;  

asset allocation is one of them,  and it may be the most important.

Asset allocation is the exercise of determining  how much of each asset class you should hold in  

your portfolio. In general, the asset classes  that we have to choose from are stocks, bonds,  

real estate investment trusts, and alternatives.  Those categories can be broken down further,  

but I will leave it there for now.  Except in hindsight there is no optimal  

asset allocation. The best that we can do is  take guidance from the academic literature.

I’m Ben Felix, Associate Portfolio  Manager at PWL Capital. In this episode  

of Common Sense Investing, I’m going  to tell you about asset allocation.

Building on the work of Harry Markowitz,  

William Sharpe and John Lintner are credited  with developing the Capital Asset Pricing Model,  

or CAPM. If you have taken any finance courses  you are at least aware that the CAPM exists. This  

was the first model that quantified the  relationship between risk and expected returns.

The CAPM is a single factor model. It sees risk  and return as being determined by a portfolio’s  

exposure to market beta, or the riskiness  of the market as a whole. Market beta is  

a priced risk. In other words, we expect a  positive outcome for maintaining exposure  

to market beta. This is very different from  betting on a specific segment of the market.  

That introduces idiosyncratic risk, which  does not have a positive expected outcome.

To eliminate idiosyncratic risk we start  our asset allocation journey with the  

market portfolio. A total market index fund  is representative of the market portfolio.

The next question is which  stock markets do we need  

exposure to? The answer, in short, is all of them.

Combining Canadian, US, International Developed,  and Emerging Markets stocks into a portfolio  

improves the risk and return characteristics  relative to each of the individual parts. This  

would be expected considering the imperfect  correlations between each asset class. US  

equities are an exception as they actually  look better on their own over many historical  

periods. We know this in hindsight,  but should not bet on it going forward.

The optimal mix between Canadian,  US, International developed,  

and emerging markets stocks in a portfolio  is an unknown. This makes the geographic  

asset allocation choice mostly arbitrary,  as long as global exposure is achieved.

Most model portfolios, including PWL Capital’s,  Wealthsimple’s, Vanguard’s asset allocation ETFs,  

and the Canadian Couch Potato portfolios have a  heavy bias toward Canadian stocks, at about ⅓ of  

the equity allocation. This might be surprising  when it is considered that Canadian stocks make  

up about 3% of the global market cap. Why so  much in Canada? It comes down to tax treatment.

I have talked in the past about unrecoverable  foreign withholding tax on foreign dividends in  

registered accounts, which is an issue that  you will never have with Canadian stocks.  

Further to that, in a taxable account  Canadian dividends receive preferential  

tax treatment. With asset allocation  being a somewhat arbitrary exercise,  

assuming some level of global diversification,  favouring a tax efficient asset class is sensible.

We have established that it is sensible to  diversify globally to attain exposure to  

the stock market, and it might make sense  to be overweight Canadian stocks relative  

to the market to reduce the impact  of both foreign and domestic taxes.

As a point for future discussion, the  Canadian dollar return for a portfolio  

equally split between Canadian, US and  International including emerging market  

stocks from January 1990 through July 2018  was 8.17% with a standard deviation of 11.92%.

Now let's talk about bonds. Bonds  are much less risky than stocks,  

and they have correspondingly  lower expected returns.

If you have noticed that Canadian bonds have  almost matched the return of Canadian stocks for  

the last 30 or so years, with less than half  the risk as measured by standard deviation,  

let me give you some context to understand  those numbers. This time period captures  

the greatest fall in interest rates in history.  Falling interest rates make bonds look amazing,  

so the numbers need to be  taken with a grain of salt.

Canadian bonds have had a low correlation to our  Canada-heavy global equity portfolio going back  

to 1990, and adding them to the portfolio looks  great. A 10% allocation to Canadian bonds barely  

reduces the return, moving it down to 8.13%, while  dropping the standard deviation more than 1%,  

to 11.83%. Keep in mind that bond returns were  uncharacteristically high over that time period,  

so we might expect a bigger drop in returns  for adding bonds going forward. Bonds would  

typically be added to a portfolio to reduce its  riskiness, not to increase its expected returns.

Finally, adding in a 6% allocation to  real estate investment trusts increases  

our historical annualized return to 8.32% while  lowering our standard deviation to 10.60%. As  

of now we have a portfolio consisting  of 10% Canadian bonds, 6% US REITs,  

28% Canadian stocks, 28% US stocks,  20% international developed stocks,  

and 8% emerging markets stocks. I have some  comments on REITs which I will save for later.

So far we have seen some big improvements by  adding in relatively small amounts of asset  

classes with imperfect correlations to stocks. I  am not going to talk about adding in alternatives  

like hedge funds, managed futures, preferred  shares, and high yield bonds. While they may  

be uncorrelated asset classes that look good in  a back test, they come with other unfavourable  

characteristics which I have discussed  in detail in past videos and blog posts.

I think that this is about as far as most people  get. Choosing some mix between stocks and bonds,  

and maybe REITs, based on their ability,  willingness, and need to take risk.

Stopping here ignores the most up to date research  on financial markets and portfolio management. In  

their 1992 paper the Cross Section of Expected  Stock Returns, Eugene Fama and Kenneth French  

summarized the body of research showing that  the CAPM has substantial shortcomings. They  

essentially concluded that the CAPM only explains  about two thirds of the return differences between  

diversified portfolios. So two portfolios with a  beta of 1 might have had substantially different  

returns, with no way to explain the difference  other than attributing it to active management.

The following year, Fama and French  proposed a new asset pricing model called  

the Fama-French Three-Factor model. Instead  of relating expected returns to market risk,  

the three-factor model relates expected  returns to exposure to the market,  

exposure to small stocks, and  exposure to value stocks. This  

model explains about 90% of the difference  in returns between diversified portfolios.

This is important for investors because if  there are three independent risk factors  

that explain returns, we want exposure to  all three, not just to one. As we have seen,  

adding in imperfectly correlated risks  should increase our expected returns and  

decrease our risk. More recent research  has identified at least one other factor  

that can be sensibly added to portfolio  construction. That factor is profitability.

These four factors, market beta, size,  value, and profitability have had low and  

in some cases negative correlations with  each other over time. The correlations of  

factors with each other are even lower than  the correlations between geographic regions  

for stocks, and in some cases lower than  the correlations between stocks and bonds.

I will quote a 2012 article published in  the Journal of Portfolio management titled  

the Death of Diversification  has Been Greatly Exaggerated:  

“ The argument that we make for factor  diversification partly rests on the  

expectation that the positive factor  premia will continue to persist. But the  

correlations (or lack thereof) these premia  with each other are at least as important”.

One of the confusing things about getting factor  exposure is that holding small cap stocks in a  

total market index fund is not sufficient. If you  have the same amount of small cap or value stocks  

as the market, you only have exposure to market  beta. It is only by increasing the exposure to  

small cap and value stocks beyond market cap  weights that factor exposure can be obtained.

Coming back to our hypothetical  portfolio, if we split up each  

geographic region into one third market,  one third value, and one third small,  

we end up with an annualized return from January  1990 through July 2018 of 9.17% with a standard  

deviation of 10.36%. Clearly adding in factor  exposure was beneficial over the time period.

This not just me cherry picking a data to  make a point either. I can’t name all of  

the studies that have demonstrated the benefits  of factor exposure, but here’s one: In a 2017  

study published in the Journal of Portfolio  Management, Louis Scott and Stefano Cavaglia  

examined the impact of factor diversification on  the odds of retirees outliving their portfolios.  

They found that terminal outcomes can be  significantly enhanced through factor exposure.

The catch is that getting factor exposure is  not always easy. It’s great if you can get it,  

but it is a challenge, especially for DIY  investors in Canada, due to a lack of products.

I said I would come back to REITs. Recent  research has demonstrated that the return  

of REITs is explained by the market beta,  size, and value factors, in addition to the  

term and credit factors which are factors that  explain fixed income returns. Based on that,  

a portfolio with exposure to the aforementioned  factors may not need an allocation to REITs.

At this point I think that total stock  market exposure is a given in the asset  

allocation decision. The mix between stocks  and bonds is more subjective based on personal  

circumstances and preferences. Exposure to  factors is important but often overlooked,  

with the asterisk that even if you  want it, it may be hard to implement.

Tell me about your asset  allocation in the comments.

Thanks for watching. My name is Ben Felix  of PWL Capital and this is Common Sense  

Investing. I will be talking about a new  common sense investing topic every two weeks,  

so subscribe and click the bell for updates.