Estimated reading time: 11 minutes
Longevity is really interesting at the moment. We thought we would take the opportunity to look back at what we did as part of the 2019 valuation, what is happening at the moment as a result of Covid-19 and what our thoughts are looking forward to the 2022 valuation.
Looking back - 2019 valuation analysis
As part of the 31 March 2019 valuation process, 20 of the English funds participating in the Local Government Pension Scheme (LGPS) advised by Barnett Waddingham agreed to take part in a project to comprehensively review their mortality assumptions. This was an increase from the uptake in 2016 which demonstrates the acknowledgement of longevity as a significant pension risk.
We considered three main analyses:
1. A detailed analysis of the mortality experience (i.e. the actual number of deaths, relative to expectations) for each fund.
2. An analysis of the socio-economic profile of each fund with comparisons so that similar funds could be grouped.
3. An analysis of the mortality improvement experience of the combined LGPS funds (i.e. by how much mortality improvement rates might change in the future).
A bit more about each of these analyses is set out below:
Mortality assumptions resulting from our analysis were set as an adjustment to the most appropriate mortality table. These adjustments resulted in a range of results with the highest and lowest life expectancies for each post-retirement membership group set out in Table 1.
Table 1: Life expectancies at age 65 by membership status resulting from 2019 mortality experience analysis of LGPS funds*
|Status||Male Lowest||Male Highest||Female Lowest||Female Highest|
*Life expectancies all as at 31 March 2019 with no allowance for future mortality improvements.
The results for some funds were combined where they had similar experience to increase the credibility of the results, as some funds experienced only a small number of deaths (with one particular fund experiencing only 17 deaths over the five year analysis period).
Generally, updates to adjustments and mortality tables, relative to the 2016 assumptions, resulted in a decrease in both life expectancy and liabilities for pensioners, all else being equal. Life expectancies varied across funds by up to 2.4 years for male pensioners but only by one year for female pensioners, suggesting that perhaps female pensioners share similar characteristics across the range of LGPS funds than males. The range was bigger for dependant members, both males and females, although there is less data for dependants than pensioners so this increase in variability of results may reflect more uncertainty in the results.
The data used for the equivalent mortality experience analyses carried out for the 2016 valuation followed shortly after funds had transitioned to the Universal Data Capture (UDC) format. This led to some instances where data was of a lower quality and data adjustments or updates were needed.
We found that the data submitted for the 2019 experience analysis was, in general, of much higher quality, and there was more consistency across the funds. It would appear that data cleansing has also resulted in a considerable improvement in data quality and therefore, in the reliability that can be placed on results.
In addition to mortality experience analysis, an analysis of the socio-economic profile of each LGPS fund was carried out using postcodes of the current membership as an indicator of socio-economic status. The analysis utilised Barnett Waddingham’s postcode tool, which was new for the 2019 analysis. It uses a member’s ‘Acorn’ type, based on their postcode, to classify the relative prosperity of different neighbourhoods where members live according to the Index of Multiple Deprivation (IMD) measure. This is a government classification ranking areas from 1 (most deprived) to 10 (least deprived).
This was useful in ensuring the robustness of our recommendations by:
Comparing pre-retirement members with post-retirement members to judge whether our recommendations for post-retirement mortality are likely to continue to be suitable for each fund in the future.
Comparing profiles between LGPS funds to understand the appropriateness of combining the experience of certain funds for greater accuracy where a low number of deaths affected the credibility of results.
Chart 1 compares the socio-economic profile of post-retirement members (pensioners and dependants combined) in each of the 20 LGPS funds who took part in our analysis. A straight line from bottom-left to top-right would indicate that the membership was drawn equally from each of the 10 IMD groups.
Chart 1: Socio-economic profile of LGPS funds for pensioners based on Barnett Waddingham postcode tool rankings by IMD.
The chart shows that members of the LGPS funds analysed come from a range of socio-economic backgrounds, but that some funds have significantly higher proportions of their membership in the lower socio-economic groups than others (the lines that are higher than the others in Chart 1). The results of this analysis tied in well with the results of the mortality experience analysis, with funds whose members live in more deprived areas generally experiencing heavier mortality (i.e. more deaths).
Based on those present, funds have not made any short term tactical investment decisions yet, as expected. Investment strategy reviews often happen after the valuation but these have been put on hold in some cases until markets settle down.
Given the strong performance over the first ten months of the 2019/20 Scheme year, diversification of assets and lower inflation the funding level has not decreased as much as you might think based on market headlines.
For the mortality experience analysis, we used the most recent version of the CMI Mortality Projections Model (the CMI Model) at the time of the 2019 valuation, CMI_2018. The model is used by the fund actuaries to make estimates about how life expectancies will change in the future. You may be aware that recent evidence has suggested there has been some slowing down in future mortality improvements (i.e. people still living longer but not increasing at such a high rate that we have seen in the past).
There are a number of parameters that feed into the model which we also considered as part of our analysis. The initial addition parameter, a new parameter, first introduced as part of CMI_2018, allows us to adjust the initial level of mortality improvements so they are appropriate to LGPS funds. A positive initial addition parameter means that we think the improvements of LGPS funds are higher (i.e. longer life expectancy) than for the general population.
Our analysis showed that there is considerable volatility in the improvements in both the LGPS funds dataset and in the general population of England & Wales for all age groups. Although there was a range of assumptions which could have been appropriate, the majority of funds adopted an initial addition of 0.5% and we are comfortable that this assumption is reasonable based on the available evidence. Using 0.5% suggests that members of LGPS funds would expect to see slightly longer life expectancies on average, compared to the general population.
"Overall the slowdown in mortality improvements resulted in a decrease in funding liabilities for most funds as life expectancies on the whole were lower in 2019 compared with 2016."
Key takeaways from the 2019 analysis
The 2019 mortality analysis resulted in all funds who took part updating their mortality assumptions to more recent mortality tables with adjustments to suit their specific mortality experience. The socio-economic analysis provided evidence to support the experience analysis, as this was consistent with socio-economic differentiators between the funds. The improvements analysis was the first of its kind for our LGPS clients and provided evidence to help funds choose how to set the new initial addition parameter in the CMI model.
The life expectancies for male and female pensioners resulting from the combined analysis are shown in Chart 2. Note that there are fewer than 20 points on the chart as some funds adopted grouped assumptions. These are higher than the life expectancies in Table 1 as they include an allowance for mortality improvements. These are the figures that are therefore used in the valuation of the overall funding position.
Chart 2: Life expectancies at 2019 valuation date for males and females at age 65 for each of the 20 funds that took part in the 2019 analysis.
Where are we now?
The Covid-19 pandemic emerged in its earliest stages in late 2019, and will not have been reflected in the longevity analyses that were used in the 2019 valuations. Barnett Waddingham is using its expert knowledge to continually analyse the effect that the Covid-19 outbreak is having on mortality for pension schemes.
Whilst we are seeing data on deaths attributed to coronavirus, the total number of deaths in the population is the important figure. Treating Covid-19 deaths as ‘excess’ mortality may not be accurate for several reasons, including but not limited to:
- Some deaths from Covid-19 may not be ’excess’ deaths, as the deceased may have died from some other cause in the same period, in the absence of the coronavirus.
- There may be some deaths from other causes that were indirectly caused by Covid-19; for example, if pressure on medical resources caused people to die who would not have died in the absence of the coronavirus.
- There may be indirect impacts on deaths due to restrictions on movement due to the coronavirus; for example, changes in traffic, pollution and mental health.
It has been reported that many of the deaths from coronavirus have been at older ages in people with pre-existing conditions, and that smokers are affected more than non-smokers. It may be the case that people who are more financially material in a pension fund (younger pensioners, and the more affluent, who tend to be in better health) are less affected by Covid-19 than the general population.
Chart 3: Weekly registered deaths since 2005, England & Wales.
Source: Barnett Waddingham calculations based on Office for National Statistics data.
Chart 3 shows the total registered deaths for England and Wales, shown by week number, in each year from 2005 to 2020 (with 2020 highlighted in orange). There has been a significant rise in all-cause deaths so far in 2020, with weeks 16 and 17 presenting more than twice the number of deaths typical for that time of year although the number of deaths has returned to normal levels in recent weeks.
There are a number of possible scenarios for how deaths could progress throughout the remainder of the year - ranging from significant future spikes in Covid-19 deaths to deaths actually being lower than average due to a large proportion of the “frail” population having died during the initial wave of the pandemic. The precise impact cannot be predicted at this stage. However it is likely that, given the size of LGPS funds, many may already have seen:
- An administrative impact - with the need to process a higher number of deaths than would usually be expected in any given year.
- A cashflow impact - with a larger number of death in service payments required in 2020 to date than would normally be expected.
- A liability impact - if there have been a larger number of deaths, the liabilities for these members will have been reduced (if a spouse’s pension is now payable rather than the member’s pension) or removed.
The impact on longevity assumptions is also uncertain, but the data that has been seen so far represents a very large spike in death rates that is unprecedented in recent years. There is much discussion in the actuarial community about the impact of Covid-19 on mortality in the longer term. The impact on your assumptions might be relatively small if the spike seen in the first half of 2020 is a one-off but factors that are yet to be seen, such as further waves, long-term health impairment of survivors of Covid-19 and the economic impacts of the pandemic could result in longer-term impact on mortality trends.
One thing that is certain is that there are interesting times ahead as further data emerges and we navigate the best way for funds to reflect the pandemic in their mortality assumptions.
Looking forward to the 2022 valuation of LGPS funds:
- We will continue to monitor the impact of Covid-19. As yet, it is too early to say what the long-term impact of the pandemic will be but by the time of the 2022 valuation, there will be much more data available to allow us to make a judgement on how best to incorporate recent experience into our assumptions.
- We are aware that the LGPS Scheme Advisory Board are taking an interest in mortality data over recent months and will be asking funds to provide death data in order for them to perform a high level analysis to estimate the impact of Covid-19 on the LGPS as a whole. Depending on the results of the analysis, the fund actuaries may be asked to do a more detailed analysis on the level of excess deaths. Either way, we would be happy to take a look at individual fund’s mortality experience since the start of the year and compare this to the assumptions made as part of the 2019 valuation. We could also break this analysis down by employer to identify which employers have been significantly affected.
- The “S3” Series tables are likely to remain the most up-to-date and relevant tables for LGPS funds. However, the ongoing analysis of LGPS mortality by the CMI SAPS Committee using data supplied by GAD will undoubtedly result in more relevant analysis to support assumption setting in 2022 and may also result in new LGPS-specific mortality tables. We will keep you updated when results are available and suggest how best to reflect the CMI’s findings in your longevity assumptions.
As new trends in longevity emerge, your Barnett Waddingham consultants will continue to keep you informed so that you can make the best decisions for your fund.
Stay up to date
Get the latest independent commentary and exclusive insights from a range of experts at the forefront of insurance, risk, pensions and investment – tailored to your preference.Subscribe today