Back to Blog feed

Duration Risk Part II

18 May 2017 by Chris Bowie

In our recent blog on Duration Risk, I argued that the structural decline in yields seen this century has been matched by a structural rise in duration; such that duration as a multiple of yield has increased dramatically, bringing increased risks to fixed income and predominantly sovereign bond investors.

We have had a lot of feedback on that blog from readers, and one interesting question we received was whether, despite the capital risks, there still remained a correlation benefit from holding long dated bonds in a multi-asset portfolio that also owned equities?

This is an interesting angle, and one we spent some time thinking about and researching. For multi-asset portfolios, there may be additional benefits from owning long dated bonds simultaneously with equities, provided the correlations are negative.

So, what is the correlation between equities and long dated bonds, and does it change over time? Using the FT Long Dated Gilt Index (>15yrs to maturity) and comparing it to the FTSE-All Share Equity Index, we find that since the turn of the century, the correlation between the two is -0.327 (using daily returns). Interestingly, we found that the negative correlation is slightly larger for FT Gilt All Stocks (all maturities) rather than specifically the long dated cohort, at -0.348 instead of -0.327.

In our view, both of these two could be described as ‘significantly negatively correlated’ over the long term, implying that long gilts (or even better, all maturity gilts) are a useful hedge against portfolios that are structurally long equity risk.

However, when you calculate rolling correlations, the picture changes significantly, as Chart 1 below shows.

Chart 1: FTSE All Share vs FT Long Gilt Rolling Correlations, Source: Bloomberg, TwentyFour

Using rolling periods that may coincide with asset allocation decisions for many investors, namely one quarter and one year, we find that correlations between the two asset classes are highly volatile – and are becoming more volatile.

Over the last year, the correlation remains negative, but less so at -0.2. On a rolling one year basis, the correlation has almost always been negative, except for a very brief period in June 2014 where correlations became positive (just). However, there have been long periods where the correlation has been lower than 0.3, such as 2005, 2006, 2007.

Perhaps most worryingly, over rolling quarters, the correlation can swing wildly, from -0.8 to +0.3. What I find most interesting is that three of the strongly positive correlations that have occurred in the last seventeen years, have happened in the last four years or so – in fact in three years out of four. Is this the emergence of a theme, or is it a feature of Q.E. driven markets?

The answer to that question is not something we will know for some time. But the short answer to the question at the beginning of the blog, is that it depends on the period over which you asset allocate. If it is more than one year, then long dated bonds should give you a negative correlation benefit against your long equity position. If you allocate over shorter periods, then you cannot take that negative correlation for granted.



This material is for information purposes only. Any views expressed are those of the author, and do not necessarily reflect the views of TwentyFour. TwentyFour does not warrant the accuracy or completeness of any information contained herein, and therefore it should not be considered as an indication of trading intent, personal investment advice, or a basis on which to buy, hold or sell any investment vehicle/instrument. As such, TwentyFour accepts no liability for any use, or misuse, of the material in this commentary. This material may not be reproduced, in part or in whole without the express prior written permission of TwentyFour.

Please remember that all investment comes with risk and positive returns are not guaranteed and you may not get back what you invested. Investing in fixed income securities comes with credit risk, default risk, inflation risk and interest rate risk.