Quant Concepts: Why diversification matters

Ian Tam: Welcome to Quant Concepts. As we near the Canadian RRSP contribution deadline of March 2^nd, I thought to take an opportunity to reiterate the importance of maintaining diversification in your portfolio, specifically to ensure a risk level that is appropriate for your risk tolerance.

At a high level, there are two main types of risks in investing, systematic risk, also known as market risk and unsystematic risk or stock-specific risk. Systematic risk is best described by what happened in the financial crisis of 2008, where the entire equity market essentially endured deep losses. Unsystematic risk and stock-specific risk occurs within a single stock or a sector and a good example of this is the general volatility of cannabis stocks in the last year or so. Systematic risk is best managed through your asset allocation or the investment between stocks, bonds, cash and other asset classes which are not directly correlated with one another. Unsystematic risk is best managed through diversifying through multiple securities within an asset class.

Today I'll use Morningstar CPMS to illustrate the effects of diversification within an equity portfolio. So, if you have a look at my screen here, I've started with a very simple model that focuses on dividends and low volatility. As usual, we're going to take some universe of stocks. Today, we're using the S&P/TSX Composite, which is about 240 companies. And we're ranking all those companies on a set of factors. The factors I'm using today are dividend yield, which is worth 25%; the historical five-year dividend growth rate, which is also worth 25%; the expected dividend growth rate, which is worth 13%. So, basically, dividends in total are worth about two-thirds of my model. I'm also going to look at the average return on equity, a measure of quality or profitability over the last five years. I'm also looking at the volatility or the consistency of return on equity over the last five years. Finally, I'm also looking at the consistency or standard deviation of earnings over the last five years. So, these six factors make up the model. And I'm going to use these six factors to rank all the stocks in the TSX Composite in my base model.

To be considered a buy a stock has to be ranked in the top 10% of the TSX Composite. To be considered a sell, this has to be ranked worse than 25% of the top quartile. So, that's what makes up my model today. And to illustrate the concept of diversification, I'm first going to run this as a very concentrated five-stock portfolio, just to give you an idea of what this return stream would look like over a very lengthy period of time. So, here I'm starting my backtest in September of 1997, picking the five stocks that best meet the requirements using the information at that point in time. Today, I'm using a $50,000 starting position in cash and are using a $20 transaction fee just to reflect a real-life situation. At the end of each subsequent month if any of the stocks break the top 25%. In terms of Frank, we sell them and immediately replace them with the next highest-ranking stock.

Looking at the results here, you can see that the strategy of five stocks, a very concentrated portfolio, returned about 8% annualized. That's net of a $20 per trade transaction cost, which beats the TSX Composite by about 1.3% and the turnover of about one-fifth of the portfolio every year.

Let's have a little closer look though at the risk statistics, specifically. So, the thing I'm going to focus on today is called downside deviation. Deviation in general looks at volatility or how consistent returns are. Downside deviation looks at the volatility when your portfolio is losing. So, just the negative volatility essentially, you'll see that over multiple trailing periods, so one-year, three-year, five-year and 10-year well, that the downside deviation is more than the index. And that's not a desirable characteristic specifically for a conservative portfolio. You typically want to look for less volatility and less downside deviation than the index that you're investing in. So not a great characteristic.

We also want to look at the number of quarters where we beat the index. In down markets, we beat the index about 65% of the time. So, it is a conservative portfolio, but it is fairly concentrated. And there are certainly ways we can improve that.

So, in the second iteration of this model, I'm going to keep everything the same, except I'm going to run this as a 10-stock portfolio. So, we're essentially doubling the number of stocks that we're investing in over that same time period. Still using a $50,000 starting portfolio or starting position, and still using $20 per trade. So, you can start to see here as the backtest finishes up just visually on the screen, the return stream is actually a lot smoother than it was with the five-stock portfolio. And that's simply because of the effect of diversification. By investing in more stocks, you're going to see that the correlation between those stocks and the interaction effect between those stocks is going to reduce the overall volatility in your portfolio and has reduced the amount of risk you're taking in the equity sleeve of your portfolio.

So, having a look at the results, you'll see not a huge difference in terms of the absolute or total return, still about 8.5% after transaction costs. But the other thing you'll notice is that the downside deviation looks like it's starting to drop down a little bit, as well as the Sharpe Ratio is a little higher than it was before. So, let's just take this one step further. I've run this test in advance. We won't look at it live on the screen. I've also run this as a 15-stock model again, adding five more stocks to the overall portfolio using the exact same parameters. I'm going to jump straight to the results today. You'll see again, the results don't change too much in terms of your total return, still, about 8.6% annualized. But most importantly, we're looking at the downside deviation and you can see here with a 15-stock portfolio in most cases, the downside deviation is lower than that index. And of course, as a result, the risk is going to be less in your portfolio, and your Sharpe Ratio is going to be higher. In this case, the Sharpe Ratio is about 0.6 over the longest time frame, which is twice that of the index.

So, again, you're getting a more risk efficient portfolio than simply buying the index. And in this case, it's even more risk efficient than simply having 5 or 10 stocks. So, in terms of the number of stocks, typically adding more stocks in your portfolio, or having a more diversified portfolio will reduce the amount of risk. The second concept I'm going to illustrate today is putting a sector limit on the portfolio. So, I'm going to take that same 15-stock model and going to run it once again. This time, I'm going to cap off the number of stocks per sector. So, what happens is, whenever a stock is sold out of the portfolio, we replace it with the next highest-ranking stock. In this particular case, if we already have two stocks in that economic sector, we're going to skip the stock and go the next one down. So, by limiting the number of stock per sector, what we're doing here is, we're ensuring we're diversified not only in terms of the number of stocks but also across the economy, ensuring that most economic sectors are represented within my portfolio.

And again, just noting, as we look at the return streams visually on the screen here, you can see it's a lot smoother. There's a lot fewer bumps over time. And again, that is the characteristics that you want, especially for a retirement portfolio or an RRSP. Where you're not really looking at taking on a ton of risk, especially if you're very close to retirement. So, here the annualized returns, about 9.6%, again after that $20 transaction cost. Sharpe Ratio is even higher about 0.7, downside deviation is about 6.3 over the longest time frame, which is certainly less than the index.

The other concept I also want to illustrate is what we call max drawdown. So, this is basically a measurement of your losses from peak to trough in terms of revenue. term profile. So, in advance I exported all these tests to show you what the max drawdown was and typically this happened around 2008 during the financial crisis. So, in the first example, the five-stock portfolio, the max drawdown was 49%. That means you would have lost half your portfolio value in the financial crisis versus buying the index, about 43% in my second test, the 10-stock portfolio, my max drawdown was 44%. So, certainly, less but still not a great situation. In my third example, 15-stock portfolio, my max drawdown was 40%. Again, a bit of it better than the index. And then, finally, if I had limited the stocks to a maximum of two per sector, the max drawdown in that final example is just 32%. So, this is another way to show you that diversifying not just in terms of the number of stocks but across sectors helps you reduce your downside risk. And this is a very clear-cut illustration of that.

Through these examples, I hope it's clear how diversifying your portfolio even within the equity sleeve of your overall asset allocation will reduce risk and provide a smoother path towards financial independence. The stocks that meet the requirements for the 15 stocks, two per sector model, will be listed in the table attached to the transcript to this video.

Click the link below to access my other video that looks at the effects of changing asset allocations for your retirement portfolio.

For Morningstar, I'm Ian Tam.