The bond market can’t be between 35.8% and 120% more risky than it used to be? Can it?
Recently I have been writing a discussion piece on our approach to the Absolute Return Sector, and one of the pieces of research I did for this was to look at the evolution of duration in the bond market since the turn of the century – and it is alarming.
Since the 31st of December 1999, the average duration of the Barclays Multiverse index has gone from 5.03 to 6.83 (as of 8th May 2017, Source: Bloomberg). Duration, or interest rate risk, has therefore risen by 35.8%.
Whilst the Barclays Multiverse is dominated by global sovereign bonds in its construction, at least it also has global Investment Grade and global High Yield bonds helping to both increase its average yield, and lower its average duration when compared to pure sovereign indices – therefore it is a good proxy for the overall investable fixed rate global bond market that many investors will allocate to. Spare a thought then for poor old conventional gilt investors, whose average duration has gone from to 6.75 to 11.62 (Source: Bloomberg) over the same period – an increase of more than 72%! Can you think of any other investment this century that has turned up the risk dial by 72% without trying? Actually, there is one; poor old linker investors, whose average duration has gone up by 120%, to the current level of 23.1 years (from December 2001 till April 2017, the longest period I could get data for, Source: Bloomberg).
This is staggering. Interest rate risk has increased by more than a third for some fixed income investors, through to having more than doubled for a different set of fixed income investors; in a time frame of less than a generation. Seventeen years in fact: less than the maturity on many sovereign bonds.
So what has driven this? Well as you can see from chart 1 below, this structural rise in duration has mirrored the structural fall in yields, so this does not appear to be just a simple terming out of debt.
Chart 1: Barclays Multiverse average yield vs average duration, Source: Barclays, TwentyFour
Of course, we have seen 100+ year issuance from several sectors (sovereigns, utilities and telecoms for example), and there was a period in the mid to late naughties where average yields rose at the same time as average duration, so it would be churlish to suggest government and corporate treasurers did not, and are not, terming out some debt. That is a perfectly rational response to the fall in yields, and with current yields near historical lows who could blame them?
So why is this duration rise worrying? Well it is not worrying if you think yields can only fall from here, because a higher duration would translate into higher capital gains for any fall in yield, compared to previously. However, we do not feel the risks on yields are one way. And that is why looking at the relationship between yield and duration in a slightly different way can be illuminating.
Chart 2: Barclays Multiverse average yield vs average duration vs duration as a multiple of yield
If you show duration as a multiple of yield (as shown above in the area chart within Chart 2), you can immediately see how expensive duration risk has become. Effectively, this measure shows you how many years’ worth of income (yield) you would lose, should yields rise by 1%. At the turn of the century, a rise of 1% on yields would lose you just under 1 years’ worth of income. An acceptable risk you might say. Today that multiple is 3.64x. This means for a 1% rise in yields, you would lose 3.64 years’ worth of income if you are invested in bonds in a similar fashion to the Barclays Multiverse index. It can be hard to visualise this, so put another way, from the date of writing this in May 2017, this will take you on a calendar basis to the end of December 2020 before you would start earning an income again. That is a frighteningly long period in which to effectively go without income payments, should your bond exposure experience a yield rise of 1%. Four Christmases without a coupon?
Remember this is for an index that has a substantial weighting to Investment Grade and High Yield credit, which improves the relationship between yield and duration significantly. If we look at a pure sovereign index such as the gilt market for example, this picture looks far worse. Currently for the gilt market, duration is 8 times the yield – implying 8 years’ worth of lost income if yields rose 100bps. On a calendar basis, that takes you to mid-2025 without a coupon.
This structural change, which has driven yields so low, and durations so high, is a risk we are actively managing against at TwentyFour. Throughout our business, in our ABS, Unconstrained and Outcome Driven businesses we are actively trying to flip the relationship on its head and ensure wherever possible the portfolio yield is greater than the duration. In our view this is the best way of protecting capital whilst still providing an excellent source of income and total return. We believe there remain many excellent sources of great total return still available within credit, but duration is not one of them. Thus across our business you will not see duration as a significant contributor to our sources of risk. Instead we believe active asset allocation, active stock selection, and a relentless focus on maximising the best breakeven opportunities in credit are the best ways of continuing to perform strongly for our clients.