The great bond switch-back
Pension funds are increasingly switching to bonds. Sometimes they are doing so for the wrong reason, with the change in fashion back to bonds, but that is another story.
Whatever your reason for buying bonds, what should you buy? When they were just a small part of a pension portfolio – 3-5% was typical of UK funds just a few years ago – then it mattered relatively little, and just buying gilts was perhaps ok. But when funds are moving to 30-40% or even 100% in the case of Boots, it matters a lot.
One notion, pushed by a few consultants, is closely to match pension payments. This is easy with the careful selection of bonds, plus the use of swaps. But it is a strange idea. Few funds really have exact knowledge of even half of future payments; the outgoings depend too much on assumptions about death, illness, marriage, employment, employer generosity etc. Precisely matching an actuary’s imprecise estimate is rather pointless. It is even more silly when a significant part of the fund remains in assets such as equities with their own imprecise payment patterns; the overall return pattern still does not match even the actuary’s imprecise estimate of outgoings.
Aside from being pointless, precise matching risks losing the extra returns available from active management of bond assets. Unlike what is often claimed about equity managers, most studies show that the majority of bond managers have been able to achieve long run out-performance against their benchmarks. Perhaps this reflects the fact that many bond holders (such as insurance funds) are indeed just asset matching, even at the cost of poor returns, thus providing active managers with inefficiencies to exploit. Perhaps recent history has not been a good representation of the long run potential of active equity managers to outperform their indices. Whatever the reason, the record is there, and it would be unwise for any fund with any risk tolerance to throw away that extra return.
It is particularly strange that many funds, when they switch from equities to bonds because of a belief – whether well or ill founded – that bonds better match their liabilities, implicitly reduce their risk budget at the same time. Why should a switch of 10 or 30% of a fund from equities to bonds lead to a switch from the effective BBB quality of equities to the AAA quality of gilts? If the switch was made in part to raise the credit quality of the return promise, then fine. But if it was made only to switch from the volatile repayment schedule of equities to the more certain schedule of bonds, then it is not clear why even an ordinary AAA rating should be required for the bonds, let alone the even higher standard of gilts.
So certainly funds should consider permitting investment managers to buy lower than AAA bonds in their portfolio. The manager might still be judged against a AAA index, such as a gilt index, with substantial freedom to buy outside that index. While equity managers are generally not permitted to buy stocks not contained in their indices, it is traditional, whatever the credit rating rules, for bond managers to be allowed, even encouraged to buy out-of-index bonds.
Alternatively, there are a variety of lower rated ‘corporate’ bond indices available for part or all of the bond portfolio. These indices are not perfect. Some are opaque – the index provider refuses to reveal what they contain. But even when their content is known, it is usually impossible or expensive to buy all the bonds contained, some of which may be very illiquid. Worse, most corporate bond indices are market cap weighted, so the most indebted issuers and sectors are the most heavily weighted. In spite of these difficulties, corporate indices and composite government/corporate indices do help managers to think rationally and their performance to be measured more objectively, so it makes sense to use them.
One interesting approach used by some bond managers, such as European Credit Management and Record Currency Management, is to buy foreign currency corporate bonds, and to hedge the currency exposure back into the domestic currency. Sometimes the duration exposure is also changed through the use of swaps, thus permitting the exact payment matching some consultants consider advisable. This approach essentially transfers one currency’s bond credit curve into another, and therefore widens the set of bonds available to the manager. It is thus better than just permitting the purchase of domestic currency corporate bonds.
But why limit the risk budget to credit risk? Surely it is better to try to add value in a variety of ways, thus diversifying the sources of risk, and reducing overall risk, or increasing the return from the overall risk taken. Surely it is better to permit bond managers to take as many types of risk, preferably uncorrelated risks, as possible.
Goldman Sachs Asset Management describes this approach as “unconstrained bond management”, though they do in fact have constraints such as maximum limits in low rated bonds. I called a similar idea “the diversified added value approach to bond management” from my time at another manager.
Whatever the name, the idea is to give managers freedom to take duration (maturity) risk different from the index, to buy semi-government (agency debt) bonds, to have different country and sector weights than the index, to buy foreign currency bonds, to take currency risk different from the index, to decide actively whether or not to hedge that currency risk (perhaps even using cross currency forwards), to buy a variety of credit ratings (preferably including below investment grade bonds, such as emerging markets and high yield), to buy asset backed securities, to buy bonds convertible into equity, and perhaps to use swaps and derivatives.
Instead of taking a large deviation in one area, duration risk for example, small deviations from the index can be taken in a large variety of areas. As many of these risks are not correlated with each other, the total risk taken is significantly smaller than the sum of the individual risks. Thus there is a much more efficient use of the available risk budget, and returns should be higher in the long run for any given level of risk agreed with the manager.
All of these considerations apply whether the bond index used is for conventional fixed interest bonds or for inflation indexed bonds. It is peculiar that funds happy to see their fixed interest managers diversify away from domestic government issues, such as gilts, often are much more restrictive with their inflation indexed bond managers. While the opportunity set with inflation linked bonds is smaller, it is still there; diversification to corporate and foreign currency credits is possible.
‘Risk budgeting’ is a fashionable term to describe something of real advantage to pension schemes: the efficient use of risk tolerance in as many ways as possible, in order to maximise the return achieved. Certainly funds should extend their risk budget to bond assets as well as equity assets.
William MacDougall is an independent investment and pensions consultant