New mortality risk model from WTW to improve longevity prediction

Global advisory, insurance and reinsurance broking firm Willis Towers Watson (WTW) has launched a new mortality risk model, which it says is the first to use medical science and the views of medical experts to improve longevity predictions and improve pricing of re/insurance.

PulseModel, as the new mortality risk model is named, brings together medical information on the impact of conditions, such as diabetes, to enable better visibility of mortality rates and patterns. WTW believes this will help insurance or reinsurance companies and pension funds to more accurately price coverage, calculate their liabilities and manage or hedge their mortality or longevity exposures.

Matthew Edwards, Head of Mortality and Longevity in Willis Towers Watson’s life insurance practice, commented; “We have been concerned for some time that the mortality models in common use do not properly incorporate medical information – such as whether people are healthy or have some disease history – quite apart from lifestyle information such as smoking status or basic medical markers.”

PulseModel was developed in conjunction with a panel of medical experts covering seven main disease groups: heart disease, diabetes, cancer, stroke, respiratory, digestive/renal, and neurological (plus common comorbidities).

The model projects annual expected transitions from healthy to disease group, as well as modelling the effect of the precise starting condition, such as the type of cancer. PulseModel also incorporates risk factors such as smoking status, blood glucose and BMI as well as the time since diagnosis, to provide more complete views of mortality risk.

WTW says that the model can quantify the scale of health related impacts to cohort mortality such as the diabetes epidemic facing the UK, while also highlighting how the spread of the epidemic can be influenced by other medical or lifestyle factors.

Dr. James Brown, Lecturer in Ageing Metabolism, Aston Research Centre for Healthy Ageing, Aston University, explained; “Type 2 diabetes is indisputably one of the gravest health issues facing us in the 21st century. Prediction of the future burden of diabetes has previously been poor, with trend-based analysis providing predictions that frequently underestimate disease burden.

“This model represents a potential step-change in our ability to accurately predict outcomes based upon the likely future trends in diabetes incidence and more importantly it provides quantifiable predictions of the potentially dramatic implications that diabetes might have on mortality and life expectancy.”

PulseModel can be used by insurance or reinsurance firms to consider future longevity trends more scientifically and to set their assumptions accordingly. With the growth of longevity risk transfer and longevity hedging, having robust models to analyse expected mortality for populations involved in these transactions is vital.

Interestingly, WTW says that the model indicates that longevity improvements will actually be lower in future, averaging somewhere around 1% annually compared to more typical current assumptions of over 1.5% per cent improvements annually.

WTW says that these perhaps inflated assumptions are largely due to regulatory pressure which “is effectively forcing insurers to overprice.”

That would suggest that pension funds and those involved in longevity reinsurance transactions may be overestimating forecasts of longevity improvement currently, making this new model likely a welcome addition to the toolbox for firms facing longevity or indeed mortality risks.

Matthew Edwards said; “What is clear from our research and the development of this new model is that we can no longer rely only on an extrapolative approach to setting longevity improvement assumptions which ignores medical views. Such approaches can give results detached from medical and biological reality and add a layer of cost which the man on the street ends up paying.”

Any enhancement to the ability to model mortality, and as a result longevity, expectations and predictions will be of interest to insurance-linked securities (ILS) fund managers that allocate to life, mortality and longevity reinsurance transactions.

Having the most accurate view possible of the risks already assumed, or that are planned to be assumed, is vital for underwriters of life insurance related risks such as these. With the claims that this model can help to make longevity predictions more accurate and medical science based, it could be very useful to those ILS funds managing life risk investment strategies.