Six Month Correlation Among iShares ETFs

Owning multiple funds with close correlations is mostly wasted effort and potentially self-deception. Owning multiple funds with a significant correlation spread can be helpful in reducing overall portfolio volatility.

Here is a table with the 6-month correlation of daily returns of selected iShares ETFs with at least 6 months of history (as of April 21, 2009), as rendered by the iShares website.

It reveals how the various funds have behaved during the past two quarters which saw so much portfolio value destruction. Longer periods of time for correlation should generally be considered, but it is interesting to see how funds behaved during these violently turbulent times.

click image to enlarge

Here is a shorter list of iShares ETFs showing their 3-year correlation:

Disclosure: We own these funds from the list HYG, MUB, TIP, LDQ, AGG,

Comments

[User 75976:]

This article is nonsensical. What are the ETFs correlated with? --- the daily production of coal in Thailand? Who knows?

Apr 23 08:30 AM

[Andreas:]

These iShares drive me really crazy. We are in a market now, it makes more sense to buy the shares one by one. Get rid of iShares. Forget them. They are a money making machine but not for you.

Apr 23 10:23 AM

[Richard Shaw:]

USER 75976 Nonsensical is a bit much. You are correct that a sentence was left out. The correlation is the the S&P 500, and I apologize for the omission.

Apr 23 12:29 PM

[Colorado St...:]

Richard,

Have you attempted to use efficient frontier methodoly in constructing an ETF only portfolio? In theory, with the diversification that well chosen ETFs across multiple sectors would create, it would seem that the efficient frontier approach would work. What are your thoughts?

Apr 23 12:40 PM

[azgm:]

Ditto with comment #1 this data is meaning less.

I have seen several articles by others that are chucked full of non sense. I wish this website's editor, would review these before posting. Otherwise they are at risk being just another bad source of info!!

Apr 23 12:45 PM

[azgm:]

Richard,
I apologize if I sounded a bit harsh too, omissions happen, I commit them everyday.

I just have seen a few to many over time on this site, having a second set of eyes before sending will usually catch most of these accidental omissions.

AZGM

Apr 23 12:46 PM

[Indexor:]

Richard:

Thanks for the data, as the last 6-month time frame has changed many of these correlations considerably. Apparently, several readers are either new to investing and not familiar with correlation tables, or simply feel it is their job to complain. In any event, I appreciate your efforts in this post and the many others you have provided here.

Apr 23 06:04 PM

[Richard Shaw:]

Indexor - thank you kindly

Apr 23 08:04 PM

[User 95663:]

What formula did you use?

Does .530 for EFA imply that half the time it is above IVV and half the time below IVV?

Thanks.


On Apr 23 12:29 PM Richard Shaw wrote:

> USER 75976 Nonsensical is a bit much. You are correct that a sentence
> was left out. The correlation is the the S&P 500, and I apologize
> for the omission.

Apr 23 10:03 PM

[No Free Cake:]

User 75976, AZGM, and User 95663, correlation is a standard investment metric. We can argue it's usefulness, but, it's a common tool. Modern Portfolio Theory (MPT) is based on it.

Even before Richard commented it was obvious the baseline was the SP500. You could see that since IVV was listed with a correlation of 1.0 in the 6-month chart. Even without that, in the US correlations are generally assumed to be the S&P unless stated otherwise.

If you follow Richard long enough you'll actually learn something. But, you do have to actually know something about investing to appreciate some of his comments.

Apr 23 10:52 PM

[Bob Mayo:]

I'm curious why you left out EEM, as many consider it one of their core equity holdings alongside IVV and EFA.

Apr 27 03:42 AM

[Richard Shaw:]

Bob Mayo, good observation. no intention in that. must have been inadvertently deleted in processing. during the week I will find that number and post as a comment. thanks.

Apr 27 07:35 AM

[Marion Cont...:]

As the author noted, longer time periods should be evaluated. In addition, different market environments should also be considered.

It would be interesting to evaluate ETF : S&P correlations during the prior bull market (i.e., for various time periods leading up to the Oct '07 peak).

Apr 27 01:47 PM