Setting a strategic asset allocation policy is an invaluable exercise for a pension fund. It is probably the most important investment decision that any pension fund makes. However, it is not concerned solely with seeking the highest return, nor with balancing risk and return. The strategic allocation is aimed at achieving the fund’s objectives.
Let’s look at the process through which pension fund trustees set an asset allocation policy and identify the benefits they derive from this process. The process is shown in figure 1.
Ideally the scheme objectives would be an input to the process rather than an aspect of the modelling. However in practice, it is difficult to set objectives without reference to the modelling.
The direct benefit from setting a strategic asset allocation is the asset allocation itself. But there is also a subsidiary benefit – setting the policy generates understanding. The understanding comes in two areas: firstly, understanding the factors affecting asset and liability values and secondly, understanding how these impact on the fund, the sponsoring company and the trustees.
This is a bit abstract. Let’s look at it in practice. The starting point in the asset allocation process is often an actuarial valuation, though it could be a major event such as the closure of a scheme to new entrants. What new understandings do the trustees gain from the asset/liability process?
The well known phrase ‘the value of investments may go down as well as up’ is undoubtedly true. However, it doesn’t say anything about by how much or when. To manage a pension fund’s assets, some assumptions for asset behaviour are needed. To make these, history, financial theory, judgement or some combination of these may be used. But whichever is chosen, the assumptions are forecasts of
the future.
It may not be necessary for the trustees themselves to make these forecasts, as their advisers can make them. However the trustees need to understand the forecasts and then be able to question them. Ultimately, the trustees need to have some faith in both the forecasts and the model.
It will be no surprise to many readers that actuarial asset values and market values have usually been different. The model for the assets will need to make allowance for this. Under the market-led actuarial basis, the asset value may be a smoothed market value. This is easier to model than the traditional basis as this still contains some actuarial smoke and mirrors.
What affects the value of the pension liability? This is really asking how the actuary values the liabilities. They project all the current and future pension payments going forward and then discount these back to give a present value. Demographic factors will play their part but, in the market-led methodology, the main influence on this calculation is the discount rate used and the inflation assumption built into the pension payments. This sounds complex but the financial markets trade a very similar instrument every day – it’s a bond. In essence, pension fund liabilities behave like bonds (whether they behave like conventional or index-linked bonds is more complex, but rests upon the inflation proofing of the pensions).
For pension fund trustees, there are three key messages from all this:
q Pension liabilities move in sympathy with bonds
q The asset allocation that best matches the ongoing liabilities is predominantly bonds
q Asset allocations away from bonds create a mismatch for which a reward should be required.
This understanding is absolutely fundamental to managing a pension fund and to making the important decision on the fund’s asset allocation policy.
The majority of UK pensions funds choose a peer group benchmark, either the CAPS median or the WM average. This is one way of defining an asset allocation policy (albeit one that the fund manager can flex and one that moves over time). An alternative, which many of the larger UK funds have chosen, is to define their own strategic asset allocation benchmark based upon their own position and objectives. The allocation chosen may be 100% in equity or 100% in gilts, but is more likely to be somewhere in between. For both of these alternatives, the key point is that asset allocation is not about seeking the highest return. It’s about trying to achieve objectives. Return is a factor, but so is the scheme’s funding level, its MFR level, the contribution rate, and so on. This is the benefit of setting an asset allocation policy. The process allows trustees to estimate and assess the various risks that they face, and in the most relevant terms.
Let’s look at a simple example. XYZ pension fund has recently had an actuarial valuation. On the back of this, it carried out an asset/liability study. XYZ initially considered three very different asset allocation policies: 100% gilts/0% equity, 65% gilts/35% equity, and 30% gilts/70% equity (with the equity allocation being a split between UK and overseas equities, and the gilts allocation being to index-linked gilts).
Using XYZ’s model, the charts in figure 2 show the scheme’s expected funding level and contribution rate in five years’ time. The columns represent 80% of the likely outcomes with mid-point representing the expected outcome.
In one respect, XYZ likes the 100% index-linked gilt option as it means that the contribution rate is very stable – there are unlikely to be many surprises. However, in another respect it can’t accept the 100% index-linked gilts option – it’s locking in a low funding level which results in unacceptably high future contributions.
The current asset allocation is 70% equity/30% bonds. Here, XYZ likes the expected contribution rate but is concerned by the range of possible outcomes – more specifically, it is concerned by the possibility of a very high contribution rate.
The next step was to conduct a great deal of sensitivity analysis. What happens if the model assumptions are changed? What happens if the actuary changes his assumptions? After all this, the trustees discussed the results with the company and the actuary. For XYZ, the final outcome was a small reduction in the equity allocation, and a small change in the split of assets between the UK and overseas. This was the allocation that best fitted the scheme’s objectives – the one that the trustees were most comfortable with.
After a process that lasted three months, the trustees fine-tuned their strategic asset allocation, but did not fundamentally change it. Were the trustees glad they went through the process of setting the strategy? Absolutely. A greater understanding of the risks of running a pension fund, and how best to meet the scheme’s objectives, was of profound importance.
Bob Cast is a consultant at Frank Russell in London

Topics