Inter-Asset Correlations

Before we move on to the next phase on investing, I want to highlight asset correlations and expand slightly from comments I made in a fixed income post. Not the easiest topic to explain nor understand. But critical in understanding why diversification plays such an important role in building a proper investment portfolio .

Diversification and Asset Correlations

We covered Diversification and Asset Correlations previously.

Correlation is a statistical measure of how one asset moves in relation to a second asset.

Many variables impact how two assets react to each other. In the linked article, I used two Starbucks as an example. At a very close level, with only a few variables, there are still some differences in operations and performance. The greater the number of variables, the greater the differences in returns of each Starbucks location.

It is an identical relationship between bankable assets, such as equities. Two Western Canadian operating oil companies’s shares will be impacted by oil prices, taxes, the economy, and so on. But differ based on their management, the ability to find productive plays, etc. This will differ too versus an oil company operating in Iraq or Texas. Because there are more variables and the impact on a Calgary versus Iraqi oil company will be different to some extent.

Correlation Coefficient

The correlation coefficient measures how closely two assets move in lockstep.

Perfect movement together is 1.0. Perfect movement in opposite directions is -1.0. No correlation between the two assets would be 0.0. Assets within a specific asset class tend to be relatively highly correlated to each other.

This figure is important when calculating portfolio risk. There is no impact on return.

If you review, Asset Correlations in Action, you can see the mechanics of how adding any asset to your portfolio that is not perfectly, positively correlated, will reduce overall portfolio risk. The level of risk reduction will be based on the correlation between the assets, as well as the quantity added.

If you have a $100,000 portfolio and add an asset with with a correlation coefficient of -0.60, you will get some nice risk impact. But if you only add $1000 of that asset, the impact is negligible. Whereas, adding $75,000 in an asset with 0.50 will actually be preferable due to the amount.

Also, while you want to look for lower correlated assets to add to your portfolio, do not forget about return. Many assets that have low correlations to stocks, for example, may also offer lower expected returns. The correlation coefficient does not reduce returns, as it does with risk. But you may suffer lower portfolio returns by adding lower return assets just to meet your goal of adding non-correlated assets to the mix.

But Do Not Exclude Lower Return Assets Outright

As I covered in Asset Allocation: Fixed Income (Part 1), that does not mean lower return fixed income should be totally excluded from one’s portfolio. Fixed income investments may provide diversification benefits with other asset classes.

Fixed income (currently) has a negative correlation to many equity subclasses. That means the assets move in relatively opposite directions. If equities fall in price, then bonds, even at lower rates, will do well. So, yes, bonds may have lower long-term expected returns than equities. But when equities experience bear markets, bonds provide a safe haven.

I say “currently”, as correlations are not stagnant. They can shift up or down (or reverse themselves) over time based on changing economic factors. Periodically review asset correlations within your portfolio to ensure they are still optimal.

For example, many years ago, emerging market equities offered excellent diversification to domestic stocks for Canadians or Americans. That is because, for many reasons, domestic companies tended to operate primarily in their own countries or regions. What was happening in Japan, Argentina, or Germany, had relatively little impact on a business serving the Mid-West US markets. But as international trade grew, the correlations also rose. Still diversification benefits to emerging markets, but less than in the past.

Going back to our Starbucks example. Starbucks began life in 1971. A single store in Seattle. In 1987, Starbucks finally expanded out of Seattle, to Vancouver and Chicago. What Australians or Swiss drank was of no concern. It was not until 1996 that Starbucks expanded outside the US, with a location in Tokyo. 25 years of revenues solely from North America. Yet by 2018, Starbucks had over 27,000 locations in 76 countries or territories. A good example of globalization and shrinking of borders business-wise. As the world changes, so too may inter-asset correlations.

Correlation of Fixed Income

Consider a few current inter-asset correlations, using data from Portfolio Visualizer as at March 15, 2019.

The current 10 year correlation between 20 Year U.S. Treasury Bonds (i.e., fixed income) and various traditional equities are: U.S. Large Cap Stocks −0.47; U.S. Mid Cap Stocks −0.46; U.S. Small Cap Stocks −0.45; International (excluding U.S) Stocks -0.43; Emerging Market Stocks -0.38.

As I wrote above, combining two assets with negative correlations is very beneficial in reducing portfolio risk. And any correlation below 1.0 (perfect positive correlation) improves overall portfolio efficiency. But the lower, all the way to -1.0 (perfect negative correlation), the better for diversification of portfolio risk.

Fixed income offers excellent diversification potential in an equity laden portfolio.

When we look at alternative asset classes, fixed income can also enhance diversification. Consider the latest 10 year correlation between 20 Year U.S. Treasury Bonds and: U.S. Real Estate −0.22; Commodities -0.22; Gold 0.17. Once again, fixed income can aid in portfolio diversification with alternative assets.

Do not forget that within the asset class itself, various subclasses may provide diversification benefits too. For example, the correlation between the 20 Year U.S. Treasury Bonds and U.S. Municipal Bonds is only 0.37. So there is still value in diversifying within the asset class.

Of course, you also need to factor in asset risk (standard deviation) and expected annual return in your portfolio calculations. Adding a high risk, low return asset to your portfolio simply because it has a low correlation may not be wise.

Note as well, going to my point on globalization impacting asset correlations. Today, emerging market equities have a very positive correlation of 0.83 to US domestic equities. 40 years ago that correlation was much lower, with much higher diversification impact. Be sure to monitor for changes in correlations over time.

Okay, enough on asset correlations. But important to understand when we begin building portfolios. Otherwise, the question is always, “Why would I ever want 3% bonds in my portfolio if stocks expect 8% annually?”