Swiss Re is a giant insurer and reinsurer that earns its money by successfully predicting and mitigating the financial consequences of natural disasters around the globe.
Last year Swiss Re launched an extensive, multidisciplined study of what it calls Life & Health. The program aims to avert a natural disaster of another sort that was set in motion by the post-World War II baby boom and that now threatens to sink Western economies.
Life & Health’s head of research and development, a new hire, is Daniel Ryan, a member of the World Economic Forum’s Global Agenda Council on Ageing, which is tackling how to meet the health care needs of an increasingly elderly population while enabling it to play a full and active part in society. Ryan also chairs the technical committee of the Life & Longevity Markets Association, a nonprofit venture that promotes a liquid, traded market in longevity and mortality risk geared toward assisting pension funds that are interested in transferring risk to the wider capital markets. Institutional Investor contributor Maureen Nevin Duffy recently spoke with Ryan.
Institutional Investor: In your paper “A Window into the Future: Understanding and Predicting Longevity,” you describe a sea change in the industry’s traditional methods of predicting mortality. What have you learned about how long we will live?
Ryan: The main difference between existing and future research is the sheer amount of data available to our department from affiliations with the medical, science, academic and other professionals we’re working with now. We have data on clinical trials as well as drug treatment influences on longevity.
What does the data say about how long we’ll live?
To start, the definition of “elderly” is 90, as defined by 70- and 80-year-olds themselves who we’ve interviewed. If we get very good at dealing with not only disease but the aging process, then we might be talking about a seven-year increase in life expectancy over the next 50 years. The average now is 78 for men, maybe 79 in the United States. So that would be an average life of 86 years for men. Something in the region of 20 to 25 percent of individuals born today might get to 100, assuming we don’t get hit with a major life-threatening disease and that individuals are prepared to take treatments while in a healthy state.
There is another school of thought that says the first person to make it to 1,000 has already been born. There are seven different ways cells can be damaged. We can deal with all seven now in terms of correcting genes within a single cell. Regular checkups could extend human life, but at a remarkably huge cost to society, with implications for resources and crowding.
To say nothing of defined benefit pension plans and social security, plans originally designed to provide a set amount until death, assuming it would occur a few years postretirement. Should defined benefit plans be discontinued across the board?
We’re trying to provide a realistic basis for assumptions. Some scenarios will have mortality lower and some higher than the pension schemes are using. Certainly, findings may be perceived as another reason to bring DB to an end, but it’s really part of a wider discussion, which is, who effectively is going to provide for individuals in their retirement? Some companies are stepping back from generous packages. We’ve seen governments, among the few still offering DB plans, trying to change retirement ages, for instance, with more success in some cases than others, such as in France, where there is clear reluctance to raise the age above 62.
You’ve written that providers might consider passing along some or all of the longevity risk to a third party. What are plan providers’ options?
There are several structures. They can do a straight transfer of assets and liabilities to a third party, akin to a plan buyout, or turn over just the DB portion (see II’s report on Prudential’s portfolio protected buy-in, “Can Insurance Companies Ease the Pension Problem?,” June 1, 2011). And they can do longevity swaps and other index-based solutions, in which the assets stay with the pension scheme, but it immunizes itself against unexpectedly high levels of longevity.
The swap can be based on an index that reflects the general population. If the plan was expecting a two-year increase and the population lives four years longer, a third party pays an amount to the pension scheme. Or the swap can be a cleaner, more complete match using an index based on the actual pension scheme experience itself. In which case, the plan would evaluate how many people have died and the payment it receives would immunize it directly for its losses.
However, the general index would be easier to market because people can see it and trade it around. If it’s more bespoke, there would be a less physical and more over-the-counter approach. Therefore providers are more reluctant to offer bespoke plans. The choice will probably depend on price, with the higher being for bespoke index swaps.
Up to now, the majority of swaps we’ve seen have used a bespoke index, with the one exception being JPMorgan, which has done a couple of swap deals using a standard population index. They’ve been able to provide a significant level of hedging, even taking into account the difference between the general population and the pension scheme population.
What should potential investors know about these capital markets solutions?
There is a difference in the counterparties involved. Some counterparties will only be interested in lining up the pension scheme with the ultimate holder of the risk. They don’t propose to hold the risk themselves. Others take on basis at some point and will pass longevity risk on to the market. Swiss Re, however, wants to hold the longevity risk. We have a substantial amount of mortality business. And long risk-taking in these deals is nicely hedged against the mortality business. So it’s a better situation than if you were just going to a reinsurer or third party who only held longevity risk. In that event you’d be intensifying the exposure within that counterparty and increasing your chances of default if for some reason longevity improvements grew even higher than expected.
You frequently refer to models in your report. Is there a financial model, a 21st-century Black-Scholes perhaps, waiting in the wings to tackle the financial uncertainty that you’ve described?
The problem with mortality modeling is different from a financial model of stock market fluctuations, because we do not have historical experience that is rich in data points and reflects the underlying volatility. Trends across the developed world have been relatively consistent for a number of decades. We have not seen in living memory a Black Death event or a sudden leap ahead in life expectancy. Consequently, we need to make use of semiquantitative approaches, where we define a scenario based on a series of future events that we think might happen and then try to quantify the impact of such events.
