Is Positive Asset Correlation A Problem?

Is Positive Asset Correlation A Problem?

An article with the provocative title “Why 60/40 Portfolios Are Downright Dangerous Today” caught my eye the other day.  It was in a financial planning trade publication targeted to financial advisors and asserted the following:

  • Bonds and stocks used to be negatively correlated with each other but no longer are.
  • Advisors are not making enough effort to manage investment risk using the bond portion of clients’ portfolios.

In this instance I can forgive the hyperbole because I agree with the latter point.  Managing bond portfolios today is considerably more complex than managing stocks, and many advisors are simplistically using broad aggregate bond index funds which are not selective enough to protect against declines due to interest rate increases or to other economic factors.

Unfortunately, the first point doesn’t hold water over the long term.  Stock-bond correlations have actually swung widely over the last hundred years and had been positively correlated for some 40 years before turning negative for several decades starting in the 1990s.

Is it a problem if correlations between these two broad asset classes are positive?  In the context of an investment portfolio, correlation is the degree to which two different investment assets move in the same direction (that is, increase or decrease) over some period of time.  If they grow or decline by exactly the same amount (i.e. 20%), they are said to be 100% positively correlated.  If one declines by 20% while the other grows by 20%, they are 100% negatively correlated.  The correlation between two assets can range anywhere from positive 100% to 0% to negative 100%.

Modern Portfolio Theory teaches us that as the correlation between two assets in an investment portfolio decreases, the volatility of the portfolio decreases as well, and vice-versa. In fact, the volatility of a portfolio is minimized when comprised of two 100% negatively-correlated assets in equal proportions.  The problem with such a portfolio is that the return would be 0%, since the growth in one asset would be cancelled out by the decline in the other.  By adjusting the proportion of each asset you could theoretically balance the risk and the return.  Take a 60/40 stock/bond portfolio (60% stocks and 40% bonds), for example.  Since stocks have higher returns than bonds, such a portfolio will grow when stocks are performing well, but not as well as an all-stock portfolio.  The real benefit occurs when stocks are declining and the bonds in the portfolio limit the losses.

The holy grail of investing is to invest in assets that will be either uncorrelated or negatively correlated with each other. Unfortunately as demonstrated above, stock-bond correlations do not remain constant over time.  But does it really matter if those correlations are not always negative?  In 1931, when large cap U.S. stocks had their worst year in history (-43%), long U.S. treasury bonds dropped by -5% (positive correlation).  In 2008, the second worst year for the S&P 500 (-37%), the same bond index was down almost -15% (again positive correlation).  Nonetheless, an allocation to bonds during both downturns would still have reduced the losses from stocks.  Even a positive correlation can help diversify a portfolio as long as it is less than 100%.

Given the transient nature of correlations, the best way to manage the risk of an investment portfolio in my experience is to diversify it across a broader scope of asset classes than simply stocks and bonds. I evaluate as many as 15 different asset classes when constructing portfolios, many of which are fixed income/bond classes that – to the point of the article above – are more likely to provide protection against rising interest rate risks.  Although we can arguably assert that the capital markets today present unique difficulties due to high stock valuations and artificially-suppressed rates, managing investment risk can be challenging during any economic environment.  Worrying about whether or not correlations are positive or negative is probably less important than properly diversifying a portfolio against the particular risks we face at any given time.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.