What’s the impact of the COVID-19 crisis on a stochastic multi-population mortality model?
On Tuesday, September 27, Professor Katrien Antonio shared her vision on this topic at a free IA|be event and networking lunch on our AG campus. More than 90 participants were present!
Assessing the impact of the COVID-19 shock on a stochastic multi-population mortality model
The COVID-19 pandemic has dramatically changed the way we live and work. For example:
- More widespread work-from-home arrangements
- The surge in e-commerce
- Changes in criteria for choosing a home (garden, home office space, etc.)
The insurance industry has not escaped the impact either, amongst others in the application of mortality tables.
The Li and Lee mortality model, a standard for projecting mortality in Belgium and the Netherlands, is calibrated on annual death counts and exposures at the level of individual ages. However, this type of mortality data are typically reported with a significant delay preventing a timely evaluation of the impact of a pandemic data point. Although other data sources that publish more frequently are available, the mortality data is then collected in age buckets.
On Tuesday 27 September at 11:30 am, Professor Katrien Antonio shared her views on the subject at a free IA|BE accredited event held at our AG Campus. She presented recent research on the impact of a pandemic data point on the calibration of a stochastic multi-population mortality projection model and its resulting projections for future mortality rates (Robben J., Antonio K. & Devriendt S., 2022, Risks).
- Stochastic multi-population mortality models allow to calibrate mortality via a common part with a country-specific deviation. As such, their advantage is that their calibration is based on a large dataset of observations from different countries.
- Remarkably, the evolution in period life expectancy of the 14 European countries with GDP per capita above the average of the Euro area shows a converging trend between 1920-2020.
- As of 2020: the IA|BE mortality model incorporates correlation between the time-dependent European trend and its Belgian country-specific deviation (from this trend) for males and females jointly.
- In the context of the COVID-19 impact analysis, professor Antonio and her team developed a technical protocol to transform weekly data registered in age buckets to annualised, age-specific observations. Evaluation of this technical protocol on 2020 Belgian and Danish data showed noticeably accurate results.
- Recent research suggests different approaches how to limit the impact (if wanted) of the by the Covid-19 pandemic affected data points. These approaches are amongst others the intervention in forecasting and weighting strategy (Robben, Antonio & Devriendt (2022)) and the intervention in model specification (KAG (2022)).
Your feedback gives us the insight to do better.
We hope that this conference has lived up to its promise… or at least that you had as much fun participating as we had organising it.
We carefully select a speaker for you that will put an innovative twist on a current topic. AG also runs monthly training sessions for its community of actuaries.
Would you like to see a particular topic on the agenda? Is there a hot topic you feel is particularly relevant? Take advantage of our surveyto share your views.
Do you have any suggestions on the organisation, the process, the catering? Let us know: we're always looking to improve your experience.
You were completely satisfied and have no suggestions for improvement? Tell us too!
We look forward to reading your comments and will take them on board for the next edition.
Oeps, de pagina die je zoekt bestaat niet meer!
Het is soms een avontuur op zich om de juiste webpagina te' vinden.
Het spijt ons je niet te kunnen tonen wat je zoekt.
