The Hidden Value of Low Correlation

We recently wrote about the reward-per-unit-of-risk ratio. Ultimately, investing is entirely concerned with generating reward-per-unit-of-risk taken. And there are two obvious ways to boost this ratio:

  1. Boost the numerator: that is, identify skilful managers who implement sound investment process with extreme discipline; and/or
  2. Reduce the denominator.

But how does one reduce the denominator? Well, there are a number of ways. One would be to consider lower-market-risk strategies, such as Montaka, for inclusion in an overall portfolio. Given Montaka’s net market exposure has averaged around 50% to date, it carries with it roughly half the market risk of a typical, fully-invested long only fund.

But there is another, more subtle way to do this. And that is by blending strategies within one’s portfolio that have low correlations with each other. Here’s why:

  • If we use the standard deviation of returns as a measure of risk for this analysis;
  • Then, by blending strategies with low correlations – even without any difference in skill level to generate returns – one can reduce their overall portfolio risk.

Here is a simple example.

Imagine we have two Portfolios, A and B. They have an identical expected return of 8%; and identical risk profile (measured by standard deviation) of 5%. But, the correlation of their return profiles to each other is only 30%.

Then, by simply constructing a New Portfolio of half Portfolio A and half Portfolio B, this New Portfolio will have a risk profile reduction of around 20%! As shown in the chart below of an actual simulation we conducted, the standard deviation of the New Portfolio is 4%, down from 5% of the two constituent portfolios.

And coming back to our reward-per-unit-of-risk ratio, if we can reduce the denominator by 20%, then this has the effect of boosting the overall ratio by 25%!

10032017_Chart 1

And this is where Montaka can potentially add significant value to client portfolios. Montaka’s returns to date have had an unusually low correlation with peer group managers, as well as with the market. This is illustrated below. Montaka’s correlation with some managers is less than 30%!

10032017_Chart 2

What is equally interesting is that the correlation between other peer group managers are significantly higher across the board. This means that blending other managers with other managers yields nowhere near the same increase in the portfolio’s overall return-per-unit-of-risk ratio.

This shines a light on key aspect of strategy evaluation. It’s not just about evaluating a strategy on a stand-alone basis. It is also about how that given strategy will complement, or not, an overarching portfolio. And low correlation within a strategy’s return profile is a hidden gem that is often overlooked.

Screen Shot 2015-11-11 at 12.08.48 pmAndrew Macken is a Portfolio Manager with Montgomery Global Investment Management. To learn more about Montaka, please call +612 7202 0100.

Leave a Comment

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.