In last month’s article, Rate of Return and Risk, I discussed how return and risk are related. One of the key points was that the types of assets that produce the highest returns also provide the greatest risk of losing principal (like stocks); another was that overall market risk can be reduced by diversifying among different investments.
Building on that theme, today we’ll look at how you can add together multiple assets whose prices are individually volatile, and yet end up with a portfolio whose overall volatility is much less than any of its components. In other words, how to create a portfolio where as far as market risk is concerned, the whole is less than the sum of the parts.
Doing this requires that we choose assets whose individual price movements are as uncorrelated with each other as possible. That absence of correlation is the key. Investing in a hundred different instruments does not reduce market risk if they all rise and fall together. The fall together part is what needs to be avoided.
Consider the following chart:
This is a hypothetical portfolio consisting of three different assets:
- SPY, the exchange-traded fund representing the S&P 500 Index (blue line)
- IWM, the Russel 200 Small-cap ETF (green line)
- EFA, the Europe-Australasia-Far East ETF. Each one of these ETFs consists of hundreds of separate stocks. (orange line)
In the lower part of the diagram is a line (magenta) representing the value of a portfolio consisting of equal amounts of each of these three ETFs, re-balanced annually, over an eleven-year period of time (late 2007 – today). This period was chosen specifically to include the crash of 2008. Our point here is to protect against market risk, and that crash is exactly the kind of risk we mean. Re-balancing the portfolio provides part of the effect we describe below. The rest was provided by dividing up the portfolio in the first place. More about re-balancing in a future article.
Note these points from the above graph:
- Although somewhat different in their movements, all three of these ETFs do represent different slices of the same asset class – stocks.
- All three usually rise and fall at the same time, if at different rates.
- Adding these three correlated assets together into a portfolio buffers the worst drops of the individual assets only somewhat. In 2008, the worst loser of the group was EFA, down over 59% peak-to-trough, with both IWM and SPY down over 52%. The combined portfolio was down about 52%. Later in the time period, some of the assets had quite notable drops of 18% to 22%, while the combined portfolio’s losses were only somewhat muted.
- The overall 11-year return, at +73%, was intermediate between the individual returns (-25% to +92%), as would be expected.
Adding together these particular assets reduced our overall market risk only minimally and we ended up with just an average return (that is, it was the average of the returns on the three individual assets).
Compare that to a different set of three assets, this time chosen because they are historically not correlated with each other.
- SPY again, representing stocks (blue line)
- TLT, the ETF that tracks the prices of long-term U.S. government bonds (green line)
- GLD, the ETF that tracks the price of gold. Stocks, bonds and gold have been known, for decades at least, to be non-correlated (orange line)
Here are the key points from this diagram:
- These three assets are very different in their movement because they represent completely different asset classes. Stocks, bonds and gold are nothing like each other. Bonds and gold are not part of what we refer to when we say the market.
- All three virtually never fell simultaneously; although each one individually had multiple major drops of anywhere from 15% to 52%.
- Adding these three non-correlated assets together into a portfolio produced a combined portfolio that buffered the worst drops of the individual assets to a very large extent. In 2008, the worst loser of the group was down 52%; but in that bloodbath for the stock market this combined portfolio dropped only 16.5%.
- The overall 11-year return, at +98%, was greater than any of the individual returns (+27% to +80%)!
This last point is definitely not what one would expect. But it happened because the value of the portfolio as a whole never fell into a deep decline. This is a consequence of the simple mathematics of avoiding large percentage losses (where a 50% drop, for example, requires a 100% gain to get even).
The ups and downs in the individual components actually damp out overall volatility – as long as those individual assets are uncorrelated.
In summary, a portfolio of assets which are uncorrelated with each other can be expected, over time, to produce both higher average returns and smaller major losses than a single volatile asset; or of a combination of assets that are correlated.
And that’s a conclusion that you can take to the bank.