Mortality forecasting in the context of non-linear past mortality trends: an evaluation
| PhD ceremony: | Ms L. Stoeldraijer |
| When: | February 07, 2019 |
| Start: | 16:15 |
| Supervisors: | prof. dr. F. (Fanny) Janssen, prof. dr. L.J.G. (Leo) van Wissen |
| Where: | Academy building RUG / Student Information & Administration |
| Faculty: | Spatial Sciences |
Mortality prognosis of Statistics Netherlands evaluated, validated and improved
Just like the explanatory factors, our knowledge and the methodology of mortality prognosis, the development of mortality rates is constantly changing. That is why flexibility is required in both making mortality prognosis and interpreting its outcomes. That is the conclusion of Lenny Stoeldraijer. She evaluated, validated and improved the mortality prognosis of Statistics Netherlands (CBS).
An accurate prediction of future mortality is becoming increasingly important due to the overall increase in life expectancy and its social consequences. The model that Stoeldraijer examines in her doctoral research is used by Statistics Netherlands (CBS) for the population prognosis. Statistics Netherlands calculates the prognosis of life expectancy. This figure is again used by the Ministry of Social Affairs to determine the future retirement age.
Stoeldraijer aims to evaluate the prediction of future mortality in the context of non-linear mortality developments in the past. Her research contributes to the debate about the degree of subjectivity in a mortality prognosis. When making mortality prognoses, different approaches and assumptions are used, in which additional, often subjective, information is used, to take account of non-linear mortality developments. A careful examination of historical mortality developments and careful consideration of the advantages and disadvantages of the additional information is therefore important.
The mortality prognosis for non-linear mortality trends can be improved by taking into account the disruptive effect of smoking on mortality trends and the use of mortality developments in other countries. The choices specifically mentioned in the methods appear to be more important than the choice of the method for the mortality prognosis. The methods used by Statistics Netherlands to include the mortality trends from other countries in the prognosis and to predict smoking-related mortality appear to be suitable. The connection of the prognosis to the recent observations has been adjusted.
The PhD research by Stoeldraijer is part of the research theme Population, Mortality and Health and the VIDI project 'Smoking, alcohol and obesity - ingredients for improved and robust mortality projections' by Fanny Janssen. Watch this video for an impression of Janssen’s VIDI-research.