New solutions to old problems
P ension professionals have long known about the existence of the duration mismatch being run in most funded pension schemes. This problem has been ignored for various reasons, including an historic lack of hedging instruments, uncertainty in valuation of long-term liabilities and buoyant equity markets.
The duration mismatch risk is simply the risk due to the duration of a pension fund’s liabilities being different from that of the assets held. The higher the duration the greater the change in value for any given interest rate movement. This means that the liabilities are far more sensitive to interest rate changes than the assets as the liabilities tend to have a higher duration than that of the assets held.
Many observers have ascribed the ‘pension crisis’ to the fall in equity values in recent years, but this was really only a part of the problem. At the same time as equities fell, bond yields decreased, so that although the bond assets increased in value, the liabilities went up at a faster rate due the greater duration. The net result – solvency positions of funds suffered badly. A harsh lesson in asset liability mismatch risk.
Why has this risk been left to run? In the context of a risk budgeting framework favoured by pension consultants this patently does not make sense. The basic principle behind risk budgeting is to maximise return for the risk taken. One should not be taking risk unless one is being rewarded for it. The pension fund is generally not expecting to receive any return for the duration mismatch risk being taken and under a risk budgeting framework such a position should be closed/hedged out. However, this has not been the case.
We have modelled the effect of reducing the duration gap on the risk of under-funding in a pension scheme in a single year. In this model the duration gap can be reduced through one of two ways: either increase the weighting of bonds at the expense of equities (moving vertically from one line to the next); or increase the duration of bonds already held in the portfolio (moving down a line from right to left) and increase the equity weighting (moving from one line to the next). Figure 1 shows the results of this simulation.
What this shows us is that risk can be reduced even though the equity weighting is increased if the duration of the bond portfolio is increased simultaneously (thereby reducing the duration gap). We can also see that reducing the duration mismatch is far more effective at diminishing risk than just reducing the equity weighting.
In the Netherlands the pension regulator, the PVK is switching to a risk based approach to measuring solvency risk. Pension funds are finding themselves heavily penalised for running an asset liability mismatch, requiring a higher solvency margin. The pension fund market is thus in the process of initiating duration hedging strategies.
In order to hedge the asset liability mismatch risk we simply need to ensure the duration of the assets better match the duration of the liabilities. There are several strategies to do this as well as a choice of financial instruments.
To fully hedge out the mismatch risk, the solution is to run a cashflow matching strategy where the pension payments projected into the future are exactly matched by the cashflows generated by the assets. This insulates the solvency position of the fund from interest rate and yield curve risk.
There is uncertainty surrounding the valuation of the liabilities. This requires assumptions on issues such as life longevity, future salary increases and inflation. The cashflow matching strategy will thus be put in place around a set of assumptions that will inevitably change requiring modification to the strategy in place.
Furthermore there is little opportunity for upside on the asset side as only the surplus can be invested in return seeking assets. The loss of the upside opportunity is offset by the elimination of most of the risk (the major risk left being that of the model – the risk that the assumptions used are wrong).
Alternatively, rather than a full cashflow matching strategy, a duration matching strategy can be implemented instead. Here the duration of the asset portfolio is increased to match the duration of the liabilities. This will protect the solvency position from small changes in interest rates and leave the asset portfolio some flexibility to seek return. The portfolio is still exposed to large changes in interest rates as well as non-parallel changes in the yield curve. One can structure the strategy to handle greater and greater interest rate movements and in doing so one gets closer and closer to a cashflow matching strategy.
Where one wishes to be within this spectrum is dependent on a number of factors: the risk appetite of the trustees, the strength of the plan sponsor and its covenant with the pension fund and the maturity of the fund. It is possible for pension funds to marry the two strategies, that is to partly cashflow match. Here the near term cashflows are matched as these can be predicted with a degree of certainly and the remaining medium to long-term liabilities are duration match, as these are more uncertain.
This results in a strategy which greatly reduces the asset liability risk of the fund yet allows for less onerous restructuring of the fund’s strategy on an ongoing basis. It allows for the retention of the investment in riskier and potentially higher return assets, something that will keep the plan sponsor happy as it makes the funding contributions less onerous.
Bonds have traditionally been the instrument of choice to match liabilities, but there are not enough bonds of sufficiently long duration. In the UK the longest government bond goes out to 2038 and the longest investment grade corporate to 2054. This is clearly insufficient when the longest liability of a pension fund can be up to 80 years.
T he alternative is to use swaps, which can be tailored to the specific requirements. Swaps are over-the-counter derivative instruments and are simply contracts to exchange future cashflows between two parties. They can be used to manage a variety of risk including interest rate and inflation. This solves the problem of lack of suitable bonds.
Figure 2 shows the movement along the spectrum of risk reduction for each of the strategies discussed above. There is also the possibility of achieving upside to better the solvency position.
How much upside is achievable is dependent on the choice of instruments used. Moving the portfolio into bonds crystallises the solvency position of the fund, which is fine if the fund is in surplus and the trustees wish to de-risk the portfolio, but is not if the fund is in deficit. On the other hand, a swap-based approach is an overlay strategy that does not require the wholesale switch out of riskier assets retaining room to manoeuvre.
One reason pension schemes have not been using swaps is that they are perceived as risky or speculative. This however is not a true reflection of the risk involved in these financial instruments. Corporations and governments use swaps and other derivatives on a daily basis to manage their risks. Pension fund trustees need to understand how a swap programme can be structured to reduce the risk in a portfolio. In reality they are no more risky than other instruments they are currently using.
It is unlikely that pension funds will be able to ignore this issue for much longer. The Dutch pension funds are already starting to act and with pending legislation soon to come into force throughout the rest of Europe, pension funds will be put under even greater scrutiny regarding how they manage risk within their schemes.
Keith Jecks is the global head of Pension Fund Coverage at ABN AMRO Bank. Weng Keong Loke is a member of the UK pensions team at ABN AMRO