Two years ago, the majority of us found ourselves in a situation we could barely have imagined even a few weeks earlier. Where working from home had previously been offered - and accepted - as a fringe benefit from an understanding employer, it was suddenly a legal requirement. One hour of exercise per day, social distancing, masks, empty roads, no travel and moribund high streets made 2020 a year to remember for all the wrong reasons.
2020 was also horrifyingly extraordinary for mortality rates. In the UK there were around 73,000 excess deaths above that expected based on mortality in 2019, according to the Continuous Mortality Investigation (CMI).
While the introduction of vaccines at the start of 2021 helped to reduce Covid-related deaths, the number was still significant (over 78,000, according to data from the UK Health Security Agency).
The pandemic has posed a significant challenge to the operations of the insurance industry globally even though, in the immediate months after the pandemic hit, a status quo remained. ‘Across the UK protection sector (life insurance, critical illness, income protection and so on) not too much has actually changed,’ says Kevin Carr, CEO of Protection Review. ‘Claims are being paid, rates haven’t shot up and products haven’t been removed.’
While claims have increased from a life insurance point of view, Carr adds that ‘most Covid-related deaths were older lives and often older than those who hold life insurance. Some may hold smaller ‘over 50s’ or funeral cover plans, but most typical life insurance policies expire around retirement age when the mortgage is paid off and the children have grown up.’
The big insurers may beg to differ with this view. For instance, Prudential reported a $2.41bn net loss in the second quarter of 2020 as opposed to a $738m net income in the second quarter of 2019.
Moving forward, it is imperative for life insurers, as they analyse their financial sustainability, to adopt an appropriate mortality model by including Covid mortality risk. But just how useful can mortality data from 2020 and 2021 be, given the abnormality of these years?
The problems start with what exactly is being used as data. While deaths attributed to Covid are one thing, the additional deaths caused by the indirect effects of Covid, and therefore the potential consequences associated with the pandemic, cannot be ignored. Government measures imposed during the pandemic may indirectly affect other causes of deaths, such as those with serious existing illnesses who didn’t or couldn’t seek medical treatment after the outbreak of Covid, thus decreasing their life expectancy. Or those who lost their jobs or income (due to self-isolation or companies going under) and the associated spike in suicide rates caused by long term unemployment and mental health issues. Even those recovering from Covid, who may have temporary or long-term kidney and liver failure, augment the decrease in survival probability among many recovered patients.
Add to this the risk of underestimation of deaths due to Covid because of under-diagnosis and limited testing, especially in early 2020. Furthermore, the 2021 mortality experience is unusual owing to factors beyond Covid deaths, including a reduction in deaths due to people having died from Covid in 2020 who might otherwise have been expected to die in 2021, known as the ‘forward displacement’ effect.
So, how has Covid’s skewing of mortality and morbidity data presented modelling challenges for insurers? Matthew Edwards is an actuary at Willis Towers Watson, where he specialises in mortality and longevity, and says: ‘The main point here is that no insurer has been able to make any substantive use of 2020 or 2021 data in thinking about either ‘base mortality’ - mortality rates ‘now’ - or mortality improvements, ie, how mortality rates will vary in the future from the assumed starting point.’
Modelling challenges
For base mortality, there are several modelling challenges. According to Edwards, the first is quantifying the forward displacement effect. ‘Covid deaths were heaviest in people with below-average health, for example, diabetics and those with prior respiratory conditions,’ he explains, ‘and so the survivors will have slightly better health. This effect is going to be negligible in the working age range, because there have been very few Covid deaths, but starts to become material at high pensioner ages.’
The second challenge is around Long Covid. There’s little reliable data here in general, but no indication yet of any material mortality impact. Insurers are generally not considering this to be a material factor.
However, the third and likely most material challenge is around endemic Covid. ‘All insurers are trying to form sensible views on the future mortality impact of Covid’s likely permanence,’ says Edwards. ‘The virus will always be with us, vaccine effects will wane, and new variants will continue to emerge. As a rule of thumb, it could well end up being like having another seasonal flu’s worth of deaths every year, in other words a few percent added to base mortality.’
Things are much harder in terms of mortality improvements, where there are a number of factors to consider. Historically, movements in health spending per capita have been a major driver of mortality improvements. The economic shock of the pandemic, and the diversion of healthcare resources to Covid, leads to a change of outlook. Many insurers are thinking some reduction in improvement assumptions is warranted.
While the NHS backlog has also contributed to an anomaly that will make calculations more difficult, and it may take up to five years for the changes to run through the system to take into account excess morbidity, Edwards does not believe that the effect will be dramatic. ‘The mortality impact is likely to be minor compared with many of the other elements. Insurers may reflect it, if at all, as a small reduction in mortality improvement assumptions.’
The challenge to data
The key issue will be what the 2022 data tells us. How predictive will 2022 be of future years? Is it indicative of the ‘new normal’, or is it still going to be wildly unpredictive, as 2020 and 2021 were?
Many insurers will be looking to split out Covid and non-Covid deaths from the 2022 data, allowing for some endemic Covid to carry on for many years, and thinking about decreases to their future mortality improvement assumptions.
Overall, perhaps the biggest challenge from an accountant’s perspective is that what was previously considered to be ‘objective fact’ – base mortality according to recent data, and future improvement assumptions based on long-standing historical trends – can no longer be treated as such. Assumptions that are supposed to be realistic ‘best estimates’ will necessarily involve material subjective adjustments.
A pandemic for all
Of course, it is not simply H&L insurers that have been impacted by Covid. In the same way that the virus itself affected people differently, so it was for insurers. The pandemic has disrupted norms across the sector.
‘It made more of a difference to cybercrime,’ says Chris Coomber, director of ERA Insurance Cost Management. ‘Claims went through the roof because everyone was told to go home and work from home but of course companies didn't put any protections in place.’ There was also a hike in claims around business interruption, for obvious reasons, while conversely far fewer claims on household insurance because home working resulted in fewer burglaries.
Covid-19 changed us in many ways and its impact has been felt, and will continue to be felt, far and wide. For life insurers in particular, the impact will continue to be felt for months and years to come. Understanding the direct and indirect effects of the pandemic on mortality rates is the vital first step that will allow insurers to better forecast their financial sustainability.