A different kind of risk
People are living longer. It may be good news for the population at large, but pension fund managers will not be celebrating. Longevity risk has become a serious issue to grapple with. Annuities are costing more, and insurers are getting agitated, demanding more information from pension funds and raising their prices. It is no wonder that one UK pension fund manager says the only solution is to “hope the members get avian flu”.
Biometric risks are defined by the EU pensions directive as risks linked to death, disability, and longevity. Across Europe, pension funds are tackling the costs of all these risks in different ways, some simply choosing to put their heads in the sand and ignore them. But for those pension funds that are taking the issue seriously, the fact that mortality rates are increasing is their biggest concern.
“We are not too concerned about the fluctuations within death and disability benefits each year. Of course, we’re concerned about a big catastrophe happening, but we’re insured against that. Because we have a lot of life annuities, we are mostly concerned with the longevity issue,” explains Tonny Anker Svendsen, chief risk officer at PKA, the Danish industry wide pension provider which manages eight different pension schemes. Svendsen points out that sometimes increases in mortality are not gradual. “What happens if we end up with a fortunate massive improvement because of a cure for cancer? It will reduce parts of our risks, but increase others.”
Pension funds complain that the cost of insuring annuities is becoming too high. Most larger schemes are self funded, but smaller schemes have little or no option but to turn to outside providers if they want to reduce their risk. Generally, the level of death and disability insurance will be determined by a number of factors, such as the level of scheme funding overall, level of benefit, and impact of a claim on a scheme’s cash flow.
Now, the move from defined benefit to defined contribution funds means that “everyone has their own individual pot of money and you can’t pack it in to pay someone else’s pension”, explains Jamie Winter, a senior consultant in the healthcare and risk consulting team at Watson Wyatt. Instead of insuring against ill health retirement packages, in the UK employees would turn, for example, to group income protection benefits. “Only something like 20% of the working population is covered by group income protection policy. It tends to be seen as complicated and by some employers as expensive. A typical cost for a white-collar group income protection scheme may be 1% or 1.5% of payroll,” Winter explains.
In Ireland, annuities cost so much that the Irish Association of Pension Funds (IAPF) is lobbying the government to create a state annuity fund. “Most pension funds in Ireland won’t buy annuities and will carry the risk themselves, because annuities are poor value,” argues Joe Byrne, chairman of the IAPF. Open market annuities are a regulatory requirement in Ireland, not an option, and the price of annuities is fixed by a limited number of insurance companies. These companies, complains the IAPF, take conservative investment decisions and conservative mortality assumptions when determining their rates.
Phe IAPF also estimates that annuities prices in the open market are up to 30% greater than the economic cost of the pension payment which they underwrite. For his part, Byrne points out that more than half the pension funds in Ireland are insolvent. “The need for insurance is even greater because they don’t have the capacity to carry the risk. In the good times, it wouldn’t be such a blow to the fund if a couple of people got sick.” He argues that if a pension fund goes insolvent, the state should provide an annuity fund rather than the private sector, because the state would not include margins.
Getting insurance can be problematic in other ways as well. If a smaller pension funds wants to insure for the costs of death-in-service lump sums, and the insurance company’s rates are quite cheap, the underwriting requirements may still be a problem. “In certain circumstances, significant evidence of good health will be required. At larger pension funds, many members can decide either
to self insure or use
an insurance company. Typically large schemes which self insure do not really want to become involved in significant medical evidence of good health requirements,” explains Richard Stroud, chief executive of The Pensions Trust in the UK.
Insurance companies say it is not their fault. Clearly, they are facing the same issues of longevity as the pension funds. “I can sympathise with pension funds’ concern about longevity risk. I know we have in the past couple of years had to make an announcement to the market about increasing our reserves for paying out annuities. We need to be absolutely sure that we have enough money to account for the fact that people are living longer,” says Simon Gadd, director of group risk at Legal & General.
Still, observers point out that not so long ago, there used to be 15 or so insurers in the market that pension funds could turn too. Now, it is proving to be extremely hard to find insurers who are prepared to take on bets.
Svendsen says the fund has been looking at options to take out protection against people living longer, be it through mortality swaps or through re-insurance companies, but that products are not widely available at the moment. But for providers that are launching products, there has been little or no enthusiasm from pension funds.
In 2004, BNP Paribas launched its longevity bond in the UK. Cash flows were based on the actual longevity experience of the English and Welsh male population aged 65 years old, and published by the Office for National Statistics. It had a maturity of 25 years.
The European Investment Bank would issue the bond, taking on all the longevity risk, with BNP Paribas acting as the arranger, structuring and placing it, and taking on some of the risk through a swap, then hedging its own position with a reinsurer called Partners Re. But it failed to generate any interest, not because of the complexity of the structuring so much as that pension funds have other things on their mind, suggests Mark Azzopardi, head of the pensions and insurance group, within the global risk solutions team at BNP Paribas.
“We are still talking to clients but not all our discussions will prove fruitful. Many pension funds are still reviewing their asset and liability positions and looking to put de-risking structures in place, whereby in addition to longevity risk, they deal with some or all of their inflation, interest risk and equity risk. Many do not want to put a longevity hedge in place until such time as they have finished the review of the pension fund,” says Azzopardi.
Some pension funds say the bond is problematic, because it makes assumptions about the overall population, rather than the specific population of a certain scheme, which could be quite different. “We don’t use derivatives on a very large scale. It’s a rather new product. Investors are wondering what the point is. If you use them with respect to your benefits, you have to assume that the population of the longevity bond is the same as your own. There is a misfit there,” says Alexander Paulis, chief actuary at the ABP pension fund in the Netherlands.
But Azzopardi believes pension funds do not have that many options. “Although the hedge is not perfect, it will be very highly correlated yet still costs less than 20% of the price of the alternatives in the market.” The alternative being insurance firms, of course.
A big problem with tackling biometric risk is the fact that mortality calculations can be so arbitrary. “The starting point is the standard mortality tables that everyone uses in the Netherlands. We make some changes to that, some adjustments based on our own data sets,” explains Niels Kortleve, manager of actuarial projects and special accounts at the PGGM scheme.
In the Netherlands, longevity risk must now be taken into account under new FTK standards. And mortality rates are changing so fast, the mortality tables being used, which are updated once every five years, are misleading. “I know some other pension plans that use old mortality tables without adjustments or improvements,” says Kortleve. And another fund manager suggests some schemes across Europe are still using tables from the 1970s.
It has significant implications when you consider the impact on liabilities. Kortleve believes that if Dutch pension funds have to make adjustments for new mortality rates, it could possibly imply an increase in liabilities of up to 10%.
And it is amazing how mortality calculations differ across the board. A recent study by the Cass Business School in London, Mortality Assumptions used in the Calculation of Company Pension Liabilities in the EU, revealed vast differences in how companies in different European countries measure life expectancy, leading to confusion over company pension liabilities.
Banking and financial services group, UBS, for example, calculated that FTSE 100 companies have a combined pension deficit of more then £40bn (e58.3bn). If this pension liability were to be calculated using German mortality tables, the deficit would become a £3bn surplus – a difference of £43bn. Using Danish mortality assumptions this surplus increases to £30bn but using French mortality assumptions the £40bn deficit becomes a £63bn deficit.
“The first thing is to do is get full disclosure so that people know what assumptions are being made. Mortality assumptions are not usually put in a report. There is no disclosure,” says Richard Verrall, head of the faculty of actuarial science and statistics at Cass. The authors of the report say they expected some variation of mortality assumptions across the EU, but that this variation is far greater than was expected or can be justified. Verrall and his colleagues have called for the inclusion of projected mortality tables to be used to calculate rates.
The lack of standardisation also has implications for the development of cross-border pensions. Verrall says it is unlikely that such developments will occur in the near future. But standardisation is not necessarily a good thing. “There shouldn’t be an industry standard for calculating longevity because each pension scheme has different demographic risk. Longevity risk is scheme specific,” argues Azzopardi. Ironically, it is the same argument that pension funds use to dismiss the notion of longevity bonds.
Still, says Verrall, some consensus must be reached. “There needs to be some sort of agreement across Europe about what’s the right thing to do about these things. They could all decide that they are not going to make any provisions for improvement of longevity, and though it’s not a wise decision, at least everybody would know where they stood.”
Until biometric risk is taken seriously across the board, there will not be a consensus. For some pension funds, it is still a case of ducking the issue, say actuaries. In Germany for example, where most pension obligations are unfunded book reserve schemes, companies have ignored biometric risks, says Raimund Rhiel, chief actuary and member of the German HR consulting board at Mercer. “It’s important that the book reserve which is calculated by the actuary is really a realistic figure. But it’s only a book reserve and that shows that most companies do not make real mortality observations with respect to their own populations.” He believes that the large bulk of unfunded book reserve schemes and those that are asset backed in contractual trust arrangements can be undervalued by up to 10% because of mortality rates.
Until very recently, biometric risk was not on any pension funds’ agenda. Now of course, things have changed. “The reason why pension funds often have problems these days with funding and solvency is not just because of the markets, but because they have been so generous on all sides,” says Georg Inderst, a pension fund consultant. “People retire often much earlier than the retirement age, and there has been generosity in terms of death benefits to relatives and widows. It is all great and wonderful, but in the good times, one forgot about liabilities and long-term structural trends. The pension promises were going up year on year. ”
That may have been true in the last two decades, but it is unlikely that pension funds will forget about liabilities anytime soon. It is even less likely that the generosity displayed with benefits in the past will continue into the next decade.