Impact of coronavirus (COVID-19) on pension scheme best estimate mortality

Executive Summary

At this point, there is insufficient data to determine the impact of the novel coronavirus COVID-19* on best estimate pension scheme mortality with reliability.

Our view as of today (10 March 2020) is:

• It is too early to predict that the impact of COVID-19 on population-level mortality in England & Wales will be material compared with normal annual variation. Our current central position on projected mortality rates for the purpose of assessing pension liabilities (i.e. not life assurance) is therefore unchanged.
• Based on discussions with insurers and reinsurers, we believe our position is broadly consistent with other longevity market participants.
• The situation could develop very quickly – we will let you know if our views, or those of the market, change significantly.

It is possible that the dominant impact on pension schemes may be the wider economic impact of COVID-19 (rather than the specific mortality impact). This aspect is not addressed by this note.

*Technically speaking, COVID-19 refers to the disease caused by the novel coronavirus, which itself is referred to as SARS-CoV-2

COVID-19 data

The death rate globally from COVID-19, as reported by the World Health Organisation on 4 March 2020, is 3.4%. This is the number of deaths from the disease (3,287 as at 5 March) divided by the number of cases (95,607). However, this figure is unreliable:

• Overstatement from bias towards severe cases. Milder cases are less likely to be picked up and reported, meaning that the reported mortality rate will tend to be overstated, i.e. higher than the true rate.
• Understatement due to delayed impact. In an ongoing epidemic, there will be people who are fatally infected, but have not yet died, meaning that the reported mortality rate will tend to be understated, i.e. lower than the true rate.
• Standardisation by age. Given that we know the impact of COVID-19 varies by age, the standard approach is to compare age-adjusted rates. However, there is, as yet, insufficient data to do this with COVID-19.

Best estimates of the true mortality rate for COVID-19 range between 0.5% and 4% of cases, with a central estimate of around 1%. By comparison, seasonal ‘flu generally kills around 0.1% of those infected.

There are major other areas of uncertainty – we don’t know:

• the current number of infected individuals either in the UK or globally,
• how quickly and easily the virus spreads from person to person – this is a critical factor in estimating its likely spread, or
• the likely impact of preventative measures or even which preventative measures will actually be employed.

It remains an open question as to whether efforts in countries other than China (typically less extreme than those taken in China) will be sufficient to stop COVID-19 from continuing to spread.

COVID-19 in the UK

There will certainly be more deaths from COVID-19 in the UK over coming weeks, both (a) as a direct result of the outbreak and (b) possibly as an indirect result of the NHS becoming overstretched (which is not unusual – mortality from all causes tends to rise during a heavy ‘flu season). However, based on the information available at present, we don’t believe that it is necessary to anticipate that COVID-19 will have a material impact on population-level mortality. At this stage, our current central position on mortality rates and projections is therefore unchanged. But, given that so much remains unknown about COVID-19, there is much more uncertainty over this year’s likely mortality than we would see in a typical year.

What a material UK event might look like

It is helpful to consider what the impact of a large-scale outbreak might have on UK mortality, even if we don’t believe this is the most likely outcome.

• If COVID-19 has a relatively large but not extreme impact on mortality rates over 2020 – e.g. 5% to 10% more deaths than in previous years – this doesn’t necessarily mean that our views on future longevity will change significantly. Initial data on COVID-19 suggests that the impact on mortality is much greater for older people and those with pre-existing medical conditions.
• This implies that a significant proportion of those dying during the outbreak would have been at most risk of dying in the following 12 to 24 months in any case, which in turn implies a healthier surviving population and hence lower mortality in the following few years. This effect would tend to dampen down the impact of any COVID-19 outbreak on pension scheme liabilities.
• If the impact is more extreme, specifically if there are a significant number of deaths of individuals who were previously (relatively) healthy during the outbreak, then it is less likely that there will be a bounce-back in mortality rates in following years. Based on current data, we view this as an unlikely scenario.

Note: the author is chair of the Continuous Mortality Investigation (CMI) Self Administered Pension Scheme (SAPS) committee and is also a member of the CMI’s Executive Committee. Any views expressed in this post are in the author’s capacity as an employee of Aon, not on behalf of CMI.