Assessing the impact of lower life expectancy

The following is an extract from our 2024 Solutions outlook.

In 2023, the Continuous Mortality Investigation (CMI) confirmed its proposal for life expectancy assumptions to fall, reflecting a belief that higher death rates in 2022 could be indicative of future mortality. This highlights a continued trend of falling life expectancy. To find out the implications for both DB schemes and markets, we studied a three-year valuation cycle, comparing 2022 with 2019.

This isn’t the first time we’ve written about declining longevity. Yet things are materially different now on two fronts. First, there has been the coronavirus pandemic and its potential influence on future mortality rates. Second, the market environment for gilts has shifted drastically.

What’s been happening to mortality?

When we look at the latest CMI model used by actuaries to project mortality improvements, we can see that there are three components that feed into mortality rates:

• The choice of base table (reflecting estimated current mortality)
• The rate of short-term improvements
• The assumed long-term rate of improvement

The model blends short-term improvements from smoothed historical mortality data with an assumed long-term improvement rate.[1]

In creating the 2022 estimates, the CMI put no weight on experience for 2020 and 2021 due to the coronavirus pandemic. But they placed a 25% weight on data for 2022 because, although mortality was higher than before the pandemic, it was less volatile and considered potentially indicative of future mortality. As you can see in the chart, the impact is a fall in life expectancies, continuing the trend seen over the past decade.[2]

Implications for DB schemes

We crunched the numbers to work out the impact of the above changes in mortality on a typical DB cashflow profile. As highlighted in the chart, there’s a modest reduction in expected cashflows.

The overall fall in the present value of liabilities is around[3] 3%, while liability hedge ratios increase with a far greater impact at longer maturities.

Longevity aside, DB pension funds have generally become better funded recently and more fully hedged. The changes in mortality tables act in the same direction, compounding these improvements.

Market impact

As the Bank of England seeks to reduce its gilt holdings, there have been question marks around who will buy all the gilts. The reduced life expectancy above reduces the overall demand for gilts by about £30 billion, exacerbating this issue, although it is likely to be ‘drip fed’ as changes in liability benchmarks take place, rather than be carried out in short order.

How material is £30bn in the context of the overall market? To put that into perspective, over the last three calendar years[4] (2021 to October 2023):

• Total index-linked gilt supply was around £70bn
• This was broadly evenly split between under 20-year maturities and over 20-year maturities
• Ultra-long index-linked gilt issuance (over 35 years) was around £13bn

The comparison is not like-for-like given that pension funds hold a mix of nominal and real bonds but it provides some sense of how this longevity-related ‘supply’ of bonds compares to historic issuance. Going forward we believe it could remain a reasonable guide as although overall issuance will likely be high, index-linked gilt issuance will likely be a low proportion.[5]

In terms of implications for markets, we believe the most obvious impact may be a modest gradual cheapening of 50-year real yields (and inflation) relative to 30-year real yields (and inflation). While there are several factors which can impact government bond dynamics, we believe this is probably the area most likely impacted by the longevity effect.[6]

The above is an extract from our 2024 Solutions outlook.

[1] For illustration, we’ve used the ‘S3’ series mortality base tables for both 2019 and 2022 and assumed the same 1.5% p.a. long-term rate of improvement – a common choice. The only difference came from the change in the short-term improvement assumption.

[2] Consistent with our analysis, this chart assumes an illustrative long-term rate of 1.5% a year and S3 mortality tables throughout.

[3] The answer depends on the valuation basis and market conditions.

[4] Source: DMO, LGIM calculations.

[5] For example, given £250bn of gross supply, 10% in index-linked gilts equates to £25bn of annual supply

[6] For example, longevity could be argued to also increase net supply at 10 years or 20 years but this would be swamped by other factors which are more important for these areas on the gilt curve.