The phrase ’self-sufficiency’ in the context of a UK-based pension scheme is not uniquely defined and is used to mean a range of things by different advisors and clients. For example, self-sufficiency is increasingly being used by some covenant advisors to suggest a funding target involving low investment risk and margins in the demographic assumptions, but without specific reserves against adverse experience.

Generally, at the heart of self-sufficiency is the idea that there is a reasonable — for example, a one-in-twenty — expectation that the pension scheme will not need to call on the sponsor for additional funding.

**What benefits should self-sufficiency cover?**

First, the benefits that a self-sufficiency funding target is required to cover must be clear. Essentially, this is a question of whether the target should cover guaranteed benefits payable on discontinuance, or whether it should be extended to cover discretionary benefits and other benefits linked to the ongoing support of the sponsor. If a scheme is to be considered in isolation of its sponsor, it is arguable that self-sufficiency should not include the benefits that require sponsor support to be provided.

The next question is whether self-sufficiency should be a funding target common to all schemes, or a scheme-specific funding target based on actual investment strategy. Schemes that have implemented a gilts-matching investment strategy may argue they are already broadly self-sufficient assuming they are fully funded on a gilts-based funding strategy. Alternatively, self-sufficiency may require a higher funding target than this, to the extent that the scheme does not invest in matching assets, to reserve against the potential downside of such a strategy. For the rest of this discussion, we will consider a common funding target to be consistent across all schemes.

If self-sufficiency is to be considered a target regardless of the decisions made on investment strategy to ultimately reach self-sufficiency, then the obvious starting point is a funding strategy that reduces as far as possible the investment risks to which the scheme is exposed. While there has been some debate in recent years, UK government gilts are still considered to be the lowest risk asset for a UK pension scheme. A gilts-matched valuation based on fixed interest and index-linked gilt yield curves therefore provides a benchmark against which a self-sufficiency target could be measured.

Financial swaps provide an alternative source of hedging in order to achieve self-sufficiency. Interest rate swaps provide a term structure for the discount rate in the same way gilt curves do. Similarly, inflation swaps can remove unwanted inflation risk. Such instruments could be used as an alternative to gilts hedging, or to supplement it.

The change from retail prices index (RPI) to consumer prices index (CPI) for statutory revaluation presents a problem for hedging inflation risk irrespective of the approach taken to hedging. A deduction could be made to RPI inflation to allow for CPI inflation revaluation where relevant, but until a market exists for hedging CPI inflation it may be appropriate to ignore the impact and to continue to fund for RPI. In addition, hedging minimum and maximum pension increases is difficult and, currently, very expensive in the swaps market. For these reasons, precise inflation hedging for pension schemes may not be practical or indeed cost-effective.

**It’s not all about investment**

Even if a scheme is judged to be self-sufficient in terms of its investment strategy, significant non-investment risks remain. Mortality risk comprises basis risk and the risk of random variations. The former is broadly the risk that current mortality assumptions, including allowance for future improvements, prove inadequate. Basis risk could be fully or partially hedged, and random variations can be expected to reduce in relative terms as the scheme size increases.

There are also other demographic assumption risks, for example assumptions regarding the proportion married and marital status, as well as operational risks such as legislative and data error risks. Any self-sufficiency target should arguably include allowance for these additional risks even though they are very difficult to predict and meaningfully value.

**Building in some investment risk**

Accepting there is uncertainty in the liability cash flow, it may be appropriate to build a small amount of investment risk into the self-sufficiency target. A financial economist might argue otherwise, but we could take an approach that anticipates modest outperformance in the assets compared to the funding target to build a buffer against adverse non-investment experience and, for example, any remaining risks resulting from a less-than-perfect hedging strategy. Adding a small amount of investment risk may not substantially increase the overall risk within the funding strategy when combined with non-investment risk, assuming the risks are largely independent.

A small sample of different funding and investment strategies has been modelled in Table 1 on a very simple basis to help illustrate and develop the above point. The modelling allows for expected demographic assumptions but assumes that demographic and operational risks are independent of investment risk and broadly equivalent to a combined annual volatility of 2.5% pa. The outcomes modelled are the probability of still being 100% funded after 10 years and the expected 80th and 95th percentile funding levels at this time.

The modelling might suggest that a gilts-matched funding target based on expected demographic experience may be appropriate for determining self-sufficiency, although a further reserve or prudence in the assumptions may be required if the likelihood of being at least 100% funded after 10 years is to approach 95%. A funding strategy equivalent to gilts plus 0.5% pa with some prudence in the assumptions could also be considered appropriate, but with a potential deterioration in the probability of being 100% funded after 10 years and increased likelihood that sponsor support will be needed.

**Additional reserves including expenses**

Two additional reserves would still be required. The first is a reserve for future expenses and Pension Protection Fund (PPF) levies, which could represent the most significant risk for small schemes. These schemes may require a disproportionately large amount of reserves to cover both expected expenses and, perhaps more importantly, uncertainty in the expenses.

A further reserve or prudence in the assumptions may also be required for random variation in mortality and other demographic experience for smaller schemes or if there is significant granularity in the scheme’s liability profile. The risk of such random fluctuations might only be diversified away by the very largest schemes.

All this probably means is that a sensible self-sufficiency target for small- to medium-sized schemes might ultimately be the solvency measure.

**Conclusion**

Self-sufficiency in its absolute form is simply not possible from a funding perspective because of demographic and non-investment risks. What then remains is a discussion regarding what is an acceptable level of risk for a self-sufficiency target. The marketplace sometimes appears flooded by different groups of advisers using their own brand of self-sufficiency that may or may not allow fully for all the residual investment and non-investment risks, or include a prudent assessment of future expenses. At the same time, interested parties such as The Pensions Regulator are placing increasing reliance on such measures.

Actuaries may therefore wish to consider a concise definition for self-sufficiency reflecting low investment risk and including some allowance for residual investment and non-investment risk and a prudent reserve for expenses and PPF levies. If an actuary uses an approach that does not allow for some or all of these items, then good practice would be to include an explanation about this and, in any event, that further support from the sponsor may be required even if it is not expected. Any weakness in a self-sufficiency measure should be made clear for anyone relying on the assessment.

______________________________________________________________

*Mark Shaw is a scheme actuary at KPMG*