Nick Sykes, European director of investments at Mercer, shows how Gilt yields are like fashion – where themes recur, but in an unpredictable way.

There has been much talk recently about Gilt yield reversion, as if there is an inevitability that Gilt yields will return to some previous state or level, based on hard economic facts. But perhaps Gilt yields are much more like fashion – examples being hairstyles or skirt lengths – in which the same themes recur, but in an unpredictable way, and with the 'zeitgeist' being the driving force, only explicable in rational terms after the event.

If Gilt yields are driven by 'hard' economic variables, as investors might like to think, then the past might be expected to throw up evidence that Gilt yields tended to revert to some sort of norm, and had a degree of predictability – because the economic variables themselves could well be forecastable. If, on the other hand, Gilt yields are more like fashion, there would be less reversion to a norm and very limited predictability. Who can say today what the fashion for hairstyles will be in five or 10 years' time?

I'm old enough to have been a student in the late 1970s and, not surprisingly, the fashions of the time – afros, bell-bottomed jeans, cheesecloth shirts – last longer in the memory than the prevailing level of Gilt yields. So here's a chart of Gilt yields since the late 1970s to the present day.

Gilt yields over the last 30 years

On this basis, fashion appears to have shown more tendency to revert to some sort of norm (afro hairstyles having made a limited reappearance of late) than Gilt yields do (or cheesecloth shirts), as Gilt yields have been in a consistent 35-year downtrend. Perhaps we can argue that the 1970s was an exceptional time, for many reasons (economic as well as psychedelic). Perhaps we need a longer perspective for understanding the behaviour of Gilt yields better.

Gilt yields since 1920

From this chart, we can see that, over the period of 1920 to the present day, Gilt yields have varied around an average of 6%. But we can also see that Gilt yields have actually spent almost no time close to this average level of 6%. This supports my theory that Gilt yields are like fashion: the 'average' male facial hair/hairstyle in the last 92 years might be 'clean-shaven/short back and sides', but this fails to capture the vast range of styles – from pencil moustaches to Mohicans – that have actually been worn over the period. On this basis, it might with confidence be predicted both that moustaches will become popular again and that Gilt yields will be materially higher than they are at present, provided a 50-year timescale is applied!

If my theory that Gilt yields are influenced by the zeitgeist, the general mood of the period – and that economics explains only after the event why Gilt yields were what they were – can we draw any parallels from history for our present-day experience?  

Well, Gilt yields were very low from the early 1930s to the late 1940s, a period of depression, emergency low interest rates, misguided policies of fiscal austerity and enormous economic and political uncertainty, followed by war, rationing, etc. I'm not predicting a re-run of these conditions and events (believing that history doesn't repeat itself exactly), but if one focuses on the mood of uncertainty, lack of confidence in the future, risk of policy error and unresolved political and economic problems, then the zeitgeist is undoubtedly one of risk aversion. This has led to keeping the money under the bed (or the financial equivalent, investing in Gilts), not going out and spending it, nor being confident in better times ahead.

So perhaps Gilt yields are really, to use the analogy of fashion again, like skirt lengths.  Currently, one could argue there is a bias towards short skirts (equivalent to low Gilt yields). Skirts could get shorter (Gilt yields could move lower), although there is a limit to how short skirts could become (in the case of fixed interest Gilt yields, the limit is approximately zero). There is also a limit to how long skirts could get (how high Gilt yields could be), although with trains, bustles, etc, this length may be more flexible than might be thought (ditto Gilt yields – circa 16% is the upper limit seen in the last 100 years but is not exactly fixed).

Our fashionable conclusion is therefore:

Skirt lengths are materially shorter than historical averages (Gilt yields are historically low) It is likely that fashion will change (not revert) and skirt lengths fall again (Gilt yields rise), but we don't know when, and the time frame may be lengthy Skirts could get shorter (Gilt yields lower) in the meantime, although there is ultimately a practical limit as to how short (how low for yields) they could become

Nick Sykes is European director of investments at Mercer