Textbook portfolio construction theory suggests that long-dated bonds, particularly sovereign bonds from the large developed economies, are the archetypal ‘safe haven’ and diversifier of equity risk. As multi asset investors, this is an important consideration. 
Equity risk is generally the largest risk we face, and so diversifying it is central to our focus on capital preservation.
Looking at the relationship between UK equities (FTSE 100) and gilts (10-year) over time, we can see that more often than not there’s a negative correlation between the two asset classes, with much of the period spent “below zero”. That said, for extended periods, the negative correlation is not particularly strong and there are a number of occasions when the correlation is close to zero, suggesting little relationship.
Source: Miton
It’s often tempting to backfill a story onto events and it’s important to remember that these two asset classes don’t exist in a vacuum. However, some of the dominant events or trends have been highlighted on the graph, helping to explain how the relationship waxes and wanes.
The 'great financial crisis' saw gilts rally, as UK equities sold off, in line with the textbook perspective, and they became increasingly negatively correlated. Then quantitative easing saw most assets rally, including equities and gilts, which temporarily saw a positive correlation, only to give way to a breakdown in the textbook relationship over the next few years as experimental monetary policy distorted behaviours.
The bubbling up of the Greek crisis triggered a reassertion of the more traditional relationship, as investors sought the safe haven of gilts, and equities suffered. Finally, Brexit saw both equities and gilts rally, with correlations moving positive. This latter example underlines that, even around shock events, the textbook relationship can’t be expected to hold.
In short, relationships are not fixed and, while we can’t know the future, it’s pretty safe to say that how gilts and equities interrelate will continue to vary over time.
In addition to our belief that we shouldn’t permanently think of gilts as a safe haven, we think there is good reason to currently question their safe haven status, a position we have held for some time. Yields are so compressed, with the 10-year gilt yield at 1.1 per cent, their capacity to provide a buffer to equities is rather limited. Additionally, our base case is for a broadly positive economic environment and for rates to move higher globally, which suggests limiting interest rate risk.
As a result, we have had low single-digit average durations across the bond portion of the multi asset range for some time, but this hasn’t always been the case. Our pragmatic approach led us to have durations in double digits across the range in the second half of 2014 – yields were considerably higher and the economic data was less positive, with inflation heading to zero back then.
As we are not index constrained, we can pick our battles within and beyond this asset class. So, not only can we be short duration within bonds but we can look elsewhere. For example, gold looks relatively more attractive as a safe haven than long-duration gilts.
This freedom is not a luxury enjoyed by bond index investors, or index-aware investors. Indeed, they face additional dangers. A dynamic not often talked about is how the average duration of bond indices has risen materially over recent years. The main driver has been the grind lower in yields, which has seen many borrowers lock in cheaper funding for the long term, thereby pushing out the average duration.
For investors who aren’t genuinely active, not only has the interest rate risk (duration) risen, but the compensation for taking that risk (the yield) has fallen. In short, from a risk/return perspective, we prefer to avoid having too much duration in our portfolios.
Anthony Rayner is manager of Miton’s multi-asset fund range. All views are his own and should not be taken as investment advice.