Swap till you drop

Hot on the heels of Norwich Union becoming the third insurer to agree a longevity swap deal, rumours abound that pension schemes will follow suit. But obstacles to market development remain. Aaron Woolner reports

Long talked about in theory, longevity swaps look poised to become a mainstream reality. Investment bank Credit Suisse has been bullishly predicting it will sign a deal with a pension scheme "within weeks" at various recent conferences - although it declined to go on record. Others are making less firm predictions. Mark Wood, chief executive of buyout firm Paternoster, confirms quoting on a dozen separate longevity swaps, but is less sure of the timetable for this to translate into deals. "It is entirely possible that a swap will be completed this year - but not this week," he says.

The backdrop to this interest in longevity swaps from pension schemes is mainly a function of current market distortions, which make not just agreeing a price, but even finalising the parameters on which price should be discussed, almost impossible. "There is a lot of pent-up demand," adds Wood. "Schemes are seeing longevity swaps as a stepping stone to a full buyout."

The attractions of longevity swaps for pension schemes are clear. Essentially, the contract is simple - cedents offload their longevity risk to a third party, which pays the expected longevity while the buyer pays the realised longevity. Collateral is posted, usually on a monthly basis, which accounts for any increase or decrease in experience. Crucially for hard-pressed pension schemes, this means there is little or no upfront cost, with the payments only occurring over the life of the swap - and, unlike interest rate or inflation swaps, their value is unlikely to swing wildly in the short term.

For Swiss Re, however, pension schemes' recent interest in longevity schemes is not a short-term issue, but a reflection of a fundamental change in the buyout market. "The bulk buyout market will plateau in the near future, and longevity insurance, especially for large schemes, will overtake it. Longevity swaps are not a flash in the pan," says Costas Yiasoumi, London-based head of pensions at Swiss Re.

Flip-flopping

Swiss Re has been accused of short-termism in the longevity market. Having originally declared it had no appetite for the risk, it executed an about-turn a couple of years back and declared an interest in the bulk annuity sector, the team for which was then scaled down massively in third-quarter 2008 to take a pure longevity swap approach.

However, Yiasoumi insists that acquiring longevity is firmly in its sights and the company is "a natural holder of longevity risk", adding that it is a "buy and hold investor in this market".

He points to the £8 billion of longevity business in the past two years as proof of Swiss Re's commitment to the sector. Although the majority of this came in the form of annuities, it also included a longevity insurance deal - essentially the same as a swap but sold only by an insurer and with premiums extending over the lifetime of the deal - with London-based insurer Friends Provident.

According to Yiasoumi, current market conditions, rather than acting as a dampener on demand, have made it far more palatable for the medium-sized and large schemes that form the main target of the new generation of buyout providers to offload longevity risk and then run-off the schemes.

"Pension plans didn't look at longevity insurance five years ago because it wasn't being marketed - but schemes have been asking for quotes since last year because pure longevity insurance is preferable in some circumstances," he says. "For example, if a scheme has a buyout shortfall that can't be closed in the short term it may aspire to a buyout. If the finances are against it, longevity de-risking may be the best short-term solution. It is easy to address interest rate and inflation risk through a liability-driven investment solution, which allows the assets to be managed in a more flexible manner."

Volatility

But Chris Holden, chairman of the Marconi trustee company that administers the company's remaining pension scheme, believes the recent bout of market volatility has made pension schemes less, not more, likely to strike a longevity swap. He says the £350 million scheme has investigated the possibility of a longevity swap, but "couldn't get the economics of it to add up". One sticking factor is price and the managing director concedes his scheme could have underestimated the longevity liabilities it is holding. In Holden's view, it is simply imprudent to commit to making payments 20 years in the future at a time when the fund is under water (and therefore unsure that it will be able to meet these payments).

David Howell, chief executive of London-based Pacific Life Re, which has an ongoing programme of longevity reinsurance, including a £1.5 billion deal in February this year (in tandem with the London office of Reinsurance Group America), also disagrees with Yiasoumi. He argues that the change in attitude to focus on longevity - rather than longevity plus asset risk - is driven by insurers and reinsurers changing.

Pointing to Bermuda-based XL Re's recent failure to offload its structured annuity book due to investor unease over the assets backing it, Howell argues that the previous approach of combining asset and longevity risk in one deal was simply unsustainable. "It used to be the assets rather than the longevity risk that were the main play - the buyer of the business was essentially saying, 'I can make more money off those assets than you can.' But this has clearly changed. The days of the £5 billion deal are gone for now, because no-one can justify that level of asset risk."

Risk appetite

But pension schemes are not the only ones interested in longevity risk transfer. Indeed, last year's £6 billion deal between Edinburgh-based insurer Standard Life and Canada Life Re dwarfed the total amount of pension scheme transfers in the market.

And it is insurers that have taken the lead in the longevity swap sector - in addition to last month's Norwich Union deal, which transferred £475 million of longevity risk out to a group of insurers, reinsurers, insurance-linked securities investors and hedge funds (Life & Pensions April 2009, page 5), there have been two further insurer longevity swaps.

Eugene Dimitriou, head of alternative risk transfer at Royal Bank of Scotland, which recently structured the Norwich Union trade, is unsurprised that insurers have transacted longevity swaps ahead of pension schemes, despite the latter's longevity exposure being far higher, at £1.4 trillion in comparison with £125 billion.

He points out that in the UK - and in fact most markets outside the Nordic region - pension schemes do not have to allocate capital against longevity risk, whereas insurers do. "Because capital requirements for pension funds are not eased with a reduction in longevity risk - as is the case with insurers - there is less of a quantifiable incentive to proceed with a longevity swap than a UK insurer," says Dimitriou.

The capital benefits of a longevity swap were confirmed by Potters Bar-based insurer Canada Life, which last year became the second insurer to agree a deal (Life & Pensions October 2008, page 6). According to John Occleshaw, executive director, life and pensions business, at Canada Life, this made perfect sense from a capital-allocation perspective. Under the UK's Individual Capital Assessment, capital for longevity risk is allocated on a best-estimate basis, to which a prudent reserve is added.

This means it is possible to construct a 'win-win' deal where the ceding insurer can offer to pay a premium on its longevity risk to a third-party investor, while at the same time receiving regulatory capital relief.

"If what we are locked into is somewhere between best-estimate and reserving level, then although we are expecting to make payments we will still get relief on our reserves. Because we can move from reserving from prudence to locking into a fixed cashflow, we don't need to hold such large reserves of capital," says Occleshaw.

The changing attitude to asset risk is mirrored by the variety of approaches to longevity seen in recent years. Occleshaw says that five years ago the number of reinsurers ready to take on standard longevity risk - different from individually underwritten impaired life policies - was "between zero and one".

Six of the best

Now, according to Occleshaw, there are as many as half a dozen players willing to discuss longevity - including Partner Re, Hannover Re, XL Re and Pacific Life Re (PLR) - but he is aware of both the limitations in capacity, even with the current level of interest, and the possibility this capacity could disappear entirely.

"There is capacity there and at the right price - but is there enough capacity to lay off all the liabilities in the insurance and pensions sector?" asks Occleshaw. "No. And there is the danger that the existing capacity could disappear at any point if any of those market participants get full up.

"As was the case with the Norwich Union deal, our main reason wasn't to get regulatory relief, but to reduce the risks on our balance sheet and explore the issues so we would have the option in future to manage our risks. This structure can be used again."

These fears over capacity are entirely justified, agrees Howell, who says that even without a change in approach to longevity on the part of insurers and reinsurers there are significant limits to how much each can take. He argues that there are two major limitations on capacity, the first being the maximum amount an entity is willing to deploy on a given transaction - "because you want to keep a deal flow. It isn't very motivational for the team if the whole year's business is completed in the first week of January".

Secondly, reinsurers' aversion to concentration risk means most would be far more comfortable holding a range of longevity exposure, rather than one solid lump. For PLR, for example, the ideal size is just £500 million, meaning that a £3 billion deal will require up to six parties, assuming PLR's appetite is reflected by other players.

And in Howell's view, the limited level of capacity means reinsurers can be picky about the books of business they consider taking on. "There is tonnes of business out there - we have a risk appetite that we don't need to compromise."

So with both Canada Life and Norwich Union publicly stating they have structures in place to roll out further longevity swaps as their risk appetite changes, is a tradable market in longevity swaps imminent? Occleshaw is not convinced. He believes there is a fundamental friction between buyers and sellers of longevity, in that buyers prefer deals structured around indexes for ease of secondary trading, whereas cedents prefer individually tailored deals that circumvent basis risk. "The problem for pension funds and insurers is that this doesn't reflect the actual experience and payments and leaves basis risk - so our deal and Norwich Union's deal were both based on individual experience."

Solutions

According to Swiss Re's Yiasoumi, one possible way of circumventing this problem is for the insurance/reinsurance to act as a warehouse of risk on deals that are structured individually and then packaged out to the capital markets using standardised indexes.

"The key to the creation of a tradable market in longevity risk is that it is standardised, so it can be easily traded. I expect that it will be based on standard indexes, so that the systemic risk is passed on but the basis risk is not," says Yiasoumi. "Even if a market develops in standard indexes it doesn't mean pension schemes will go down that route - instead, the insurers will provide bespoke solutions and then use an index to offload the risk on to the capital markets."

In Howell's view, this would conform to the traditional approach to insurance securitisation, which is to cover events that are well out-of-the-money, such as earthquake risk or global pandemics. He argues that the longevity swap market will develop in a way that sees longevity risk parcelled off into three separate tranches: the base layer, which will kept by the insurer; the middle layer of risk, held by the reinsurer; and the tail risk. The last of these will be moved on to the capital markets, because while reinsurers are used to taking the volatility via the income statement through changes in the mortality experience, investors require a greater stability.

"Why do reinsurers retain the median risk?" asks Howell. "Investors on the capital markets like products that are highly rated and (supposedly) safe - just look at the default probabilities for an A-rated security.

"Many investors are limited to BBB or stronger - someone who wants to invest in a bond aims to make money 95 times out of 100. Any longevity risk has to be out-of-the-money so as to achieve a high enough rating to attract these investors."

Lack of information

The importance that investors place on reducing volatility is supported by Andrea Cavalleri, head of life origination and structuring at London-based insurance-focused hedge fund Securis Investment Partners, who says it is this lack of information that acts as a major barrier to investment - and competitive pricing. Securis has been looking to take on a longevity investment for two years, but Cavalleri says it has not yet found a "compelling investment opportunity".

He says the current collateralisation of swap contracts, whereby only the deviation in the block's mortality experience requires a posting of collateral, makes it a highly leveraged trade for an investor.

"So if an insurer moved from medium to long cohort it would result in a difference of 5% to the total liabilities. Because the only collateral posted would represent a minor fraction of total liabilities, this swing of 5% becomes amplified by such a factor that it leads to a potential swing in the size of your investment of even more than 100%."

RBS's Dimitriou looks towards the mortgage securitisation market for an analogy of how to get disparate books of business on to the capital markets: "In this sense, a longevity swap is similar to other types of securitisations. Take mortgages - residential mortgages for UK and Dutch banks have similar benchmarks, even if the underlying components are not similar. Indeed, if you look at the UK-specific market, how do you compare buy-to-let landlords with private loans originated at high-street banks?"

Mortgage securitisation may be an odd allusion for Dimitriou to use to talk up a new form of capital market risk transfer, but he says predicting future life expectancy is intrinsically more complex than the bricks and mortar underpinning mortgage securitisation.

A young market

"Longevity is an international risk and less open to manipulation by market players than other, more mainstream, risks," says Dimitriou. "But it is still in its infancy. When interest rate swaps started we didn't see 60-years swaps from the first day - it started short and got longer. We're now in the price discovery stage."

But according to Cavalleri, this still does not solve one basic problem - price. "The recent Aviva trade is 10 years - but other trades are 40 years or more, so having information on past mortality experience for the block only on the last 10 years of trades just isn't enough. We have failed to find out why best estimates should be based on medium cohort or average of medium and long cohort (albeit with a floor). This is the biggest single disincentive," he says.

"The reason sellers want to offload at this price is that the capital requirements on their remaining books of business are held at such unrealistic cohort asumptions," he adds. "The capital markets are not there to pick up the tab for insurers and pension schemes."