Counting the cost of improving lives
We are all living too long. Both the UK Pensions Regulator and the Actuarial Profession, the trade body for Actuaries, have recently issued warnings on this. But first the good news. booming stock markets up until the recent market turbulence shrunk the aggregate value of defined benefit pension deficits from £66bn (€95bn) a year ago to around £31bn.
This deficit is shared among about half of the 7,800 schemes measured by the Pension Protection Fund's (PPF) 7,800 Index. Every defined benefit (DB) scheme is obliged to pay a per member annual levy, part flat-rate, part risk-adjusted to the PPF. The PPF's Section 178 definition of surplus is stringent and is based on the cost of buying out all PPF benefits via the purchase of annuities. "The importance of the PPF selecting S178 is that it says how much a scheme has to hold to provide the same certainty as an insurance company can though pension annuities," judges Mick Moloney, head of European financial strategy at Mercer Human Resource Consulting.
By the less stringent accounting standard FRS17, many more schemes are now in surplus. All this should bring relief to worried trustees, but as one set of risks is squeezed back in the box another pops out. "Increasing longevity is the new, uncontrollable form of risk threatening to disrupt scheme risk budgets," adds Moloney.
The Actuarial Profession is engaged in the ongoing study of our mortality expectations and has issued recent warnings that the expected rate of improvement, running at 2-4% per year, may be too low. This means that schemes will have to pay pensions longer than expected and adjust their measure of portfolio risk accordingly. "More and more FTSE 100 companies are altering the assumed rate of mortality improvement they employ to value future liabilities," warns John Belgrove, a senior consultant at Hewitt Associates, "from the so-called short cohort to the medium cohort."
One consequence to flow from this is more attention to the granular detail of scheme specific longevity risk. This can be a more exact science, depending on the quality of data available. "Scheme size is a real issue," says Richard Giles, an actuary and head of longevity consulting at Mercer Human Resource Consulting. "We need sufficient clean data to comprise a meaningful analysis."
At least 30,000 life years are likely to be required; this could be built from 10,000 members each in a scheme for three years, or 5,000 for six years and so on. Smaller schemes may not be able to present sufficient data for this purpose. The more data, the more confident actuaries can be when they out this into a stochastic model of possible longevity outcomes for the scheme.
And while they rely partly on data from the UK Office of National Statistics, much of what is required comes only from their client schemes.
Having proprietary data of this kind as a background to informed judgement is vital when the data available from a particular scheme is not sufficient. Trustees need to test the adequacy of the advice they receive on this.
But even after these studies, trustees retain the fully responsibility for making any final decision about the rate of assumed longevity of their scheme when they are constructing the scheme specific investment strategy. "We simply cannot predict the possible effects of medical science on life expectancy," adds Giles, "and if you doubt this consider the possible consequences of the universal availability of a medicine like statins to men over fifty."
There is now real debate over issues such as the possible consequences of synthetically manufactured organs on life expectancies. Aside from looking at medical histories, life insurance providers are also taking account of individuals' postal codes when estimating their life expectancy. Occupational history, and income levels are also vital predictors of life expectancy. They are strongly correlated with access to medical care and certain forms of risk taking. Insurers' attention to this detail is becoming more granular with every year and this granularity is now finding its way into the scheme mortality estimates being composed by actuaries.
At present, there are few tools available to hedge out this increasing mortality risk although a number of investment banks say they are they developing swaps that will do so. These will be similar to the market, inflation and interest rate swaps already available to trustees. Using swaps in this way is part of the liability driven investment(LDI) that is recasting much established practice on trustee's approach to risk management.
To understand the extent of this recasting it is worth recalling that 10 years ago there was still active debate over whether trustees ought best to measure long only manager performance on a peer group basis or relative to a tradable index like the FTSE All Share.
Cash or LIBOR plus absolute return benchmarks were virtually unknown but now these are rapidly superseding all relative performance measures. The reason is simple. Schemes running swap programmes as part of an LDI investment strategy have to pay the banks providing the swaps annual cash premiums to maintain the swap programme. This renders the cost of meeting those risks eliminated by the swaps into cash rather than an assumed equity/bond rate of return.
LDI is an ill-defined concept but it always hinges on using a stochastic projection of scheme's future liability stream as a return benchmark for the scheme's investment portfolio. Risk can be measured in many ways but here it is as relative between the scheme's liabilities and portfolio of assets. LDI is implemented through a scheme specific risk reducing swap overlay programme which requires an optimally diversified asset portfolio to meet its cost and if possible to generate growth in excess of this cash requirement.
Strategies like LDI are rarely implemented in one step but via an iterative process between trustees, the scheme sponsor, advisers and the industry regulator. So what we are seeing at present is the middle of a process of transition from the old strategic consensus to the new, LDI-based one.
Mismatches between the expected portfolio return and the liability stream need compensation, or should be reduced. If not they are likely to result in scheme top-ups from sponsors, which the latter are keen to avoid. Much here depends on the attitude of the trustees to risk, and the strength both of the sponsors covenant and of their balance sheet position.
For instance, a very strong covenant from an employer easily capable of topping up any scheme deficits permits far more equity risk in the portfolio than a weak covenant from a weak employer for the same set of liabilities. Swaps allow much of this mismatch to be eliminated but paying their cost requires optimal diversification in the portfolio between sources of uncorrelated alpha.
This is why trustees and their investment consultants have been searching for new, uncorrelated asset classes and investment strategies. Examples here include infrastructure investment, actively managed currency, passive and active commodity exposure, private equity, sub-investment grade and emerging bonds and hedge funds. Ten years ago most even large UK pension funds would not have invested into any of these with the possible exception of some exposure to private equity.
Meanwhile there is also plenty of balance sheet re-engineering taking place to meet current and future liabilities as part of LDI strategy via the provision of contingent assets. Examples include Sainsbury's recent securitisation of their £8.6bn property portfolio. Tesco and Marks & Spencer have entered into sale and leaseback arrangements with their retail real estate in order to fund their pension liabilities.
"Many schemes in current surplus are using their spare cash to buy the protection offered by LDI strategies," observes Belgrove. Doing this inevitably means selling away future upside potential for downside protection and this could yet prove to be a very costly exercise.
Market timing may prove to be important determining the long-term cost of this protection. Post Myners, and a raft of legislation, trustees are expected to know their investments and understand risk, or at least to get qualified advice on both.
This is an inherently complex subject; the PPF has recently put on hold their attempts to include scheme specific measures of investment risk when setting the risk-adjusted levy. This is an indicator of how hard it is to make valid predictions about liabilities of defined benefit pension schemes.
At the heart of that of difficulty lies the issue of future mortality rates.