Cowley: I'm announcing my retirement...from long-duration bond management

I am announcing my retirement. Maybe it’s the time of year or maybe it’s my time of life, but this really isn’t worth it anymore. Doing this job has become more than difficult – it’s become impossible to guarantee that, given we are all fallible human beings and not magical prodigies, we can put a positive return on the table for our clients for anything more than a passing moment in time. And even if we do, it might just be because we got lucky.

You can come to this conclusion from a number of different points of view which include economics, sociology, politics or psychology (or any permutation of those four) but there is (and please don’t switch off just yet when I say this) a mathematical reason for it. To put it nicely; as fallible human beings doing our best in an increasingly difficult set of circumstances, our chances of guessing every trick and turn of the markets has declined to a vanishingly small number. You WILL make a mistake and it’s going to cost you or your clients when it happens. The only thing in question is how much you will lose.

Loss of income buffer

Of course, nobody gets it right all of the time – that’s not the point. The point is that there was a time when you could make a mistake and you could sort of get away with it. The margin for error was just that much larger. If you look back on the history of yields you will see (the US is just one example) that bond yields have fallen from about 14% in the early 1980’s to eye-wateringly low levels today; a five year maturity bond in the US now yields about 0.7% whilst a 30 year maturity bond yields just under 3%. In turn the amount of income generated has fallen whilst at the same time the price sensitivity of bonds to yield movements has risen (the so-called ‘duration’ has increased). So whereas once upon a time there was plenty of face-saving income being generated if you got the direction of market wrong, that has now declined. A lot.

It’s possible to calculate how much of a mistake you could make and get away with it; simply ask the question “What is the yield rise that causes the price of the bond to decline by the same amount as the income I would earn in a year such that I break-even”. You can get the answer pretty easily. If you do this for a range of maturity of bonds the increasingly dangerous mathematical situation reveals itself all too easily.

In the early 1980’s you could misjudge the direction of yields by between 2% and 4%, depending on the maturity of the bond, and no harm would be done to the absolute value of your clients’ money over a one year horizon. But now that figure has declined to below 0.15% for all bonds.  In other words, once upon a time your margin for error was vast but now it is tiny – if yields rise by more than 0.15% then even over a one year holding period you will start making a loss in absolute terms. Even demanding clients will accept that there are limits to how accurately any fund manager can predict the future, but now the level of pin-point accuracy required has contracted to ridiculous levels.

Worse to follow 

But there is more bad news to come. The added problem is that the volatility of yields has also risen. So now not only has the margin of error declined but also the probability that you will be right has proportionally declined. In effect we have been gradually losing control over our ability to produce positive returns even when we make a mistake. Just for fun, let’s set the moment when we lost control of what we were doing for a living as the time when yield volatility exceeded our margin for error – that is, in slightly more technical language, when the standard deviation of yield movements became greater than the break-even yield movement.

If you do that, then, for spurious accuracy, it’s possible to say that in the summer of 2010, as bond managers, we entered into a Never-Neverland where randomness exceeded the protection afforded by the markets if we got things wrong. We have been living on borrowed time for two years now.

This should be a sobering thought for investors in government bonds. At the very least nobody should be messing around with government bonds with maturities greater than about 5 years without lots of built-in protection. But more broadly, and with that in mind, I am announcing my retirement – from long-duration bond management. Yields may go down from time to time in the future and if a bond manager you own captures that then all well and good and there may be some genuine skill involved in that decision over the short-term.

But, for me, the thrill has gone because, now, the consequences of being wrong and the probability of being wrong are too great to reliably invest our client’s money in a sensible way to create systematic positive returns. I’m not ruling out the odd occasional long bond investment but that is what it will be – occasional, opportunistic and fleeting. Of course, there is another announcement to come; my new career as a bond manager who is habitually short the bond markets, but that has yet to come.

Stewart Cowley is manager of the Old Mutual Global Strategic Bond fund.