Option FanaticOptions, stock, futures, and system trading, backtesting, money management, and much more!

Correlation Confound (Part 2)

In the last post, I defined both words in the title. Today I continue by describing the correlation confound of portfolio diversification.

Combining assets with low correlations in a portfolio may allow me to get more return while taking on the same level of risk. It may also allow me to get the same return with less risk. This is diversification.

Risk, or variability of returns, is what causes people to close positions for the worst possible losses.  Averaging +10% per year is great for a portfolio but if, at some point during the year, you were down 80% then would you still be in the market? In 2012 I described this scenario in terms of maximum adverse excursion. Diversification helps to lower risk and while that may lower returns as well, if it can keep me in the trade mentally then time has repeatedly been shown to work its magic and allow the market to rebound.

To build a diversified portfolio, we are advised to look for assets whose returns have not historically moved in the same direction. What I do not see in most of these discussions is the fact that correlation can change.  Jim Fink addressed this in a 2013 article:

> A large portion of the disappointment can be traced to
> the severe bear markets… when correlations among asset
> classes increased markedly at the worst possible time,
> resulting in all declining in price at the same time.
> [Mebane] Faber uses 2008 as a prime example:
>

>   "The normal benefits of diversification 
>    disappeared as many non-correlated asset 
>    classes experienced large declines 
>    simultaneously. Commodities, REITs, and 
>    foreign stock indices all suffered 
>    drawdowns over 50%."

>
> If only there were a way to avoid exposure to risk assets
> during the most severe bear markets, the problem of
> converging correlations could be avoided and the
> diversification benefits of different asset classes with
> normally low correlations could be fully realized . . .

Correlation confound #1 is changing correlations. If this happens then your best efforts to diversify and minimize losses may not be effective.

Comments (3)

[…] confound #1 dealt with portfolio diversification.  Today I will discuss correlation confound #2, which has to do with pairs […]

[…] confound #1:  correlations based on historical price changes can and will change in the future.   The industry advertises diversification as having the potential to improve returns for a […]

[…] correlation confound #1 where historical values change in the future, risk tolerance can also change.  How many investors […]

Leave a Reply

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