Modelling The Risk Of Mortality In Romania

In the last 3 centuries, researchers from different area of expertise, such as, demographers, medical doctors and actuarial mathematicians, have been struggling to develop a better model to estimate biometric functions. Such a model is long due in order to improve the present methodology regarding certain statistic-demographic rates. The issue of missing data (for older ages), the issue of computing correctly the average expected life and last but not least the forecast of mortality, could be solved through the use of better models that can determine the components of life tables. A life table represents a means of determining the probabilities of an individual living to or dying at a certain age. A better image of the ageing process for human population is shown by determining the probability of death occurring at specific ages over specified periods of time. Parametric models for the projection of mortality rates were first introduced by Lee and Carter (1992) in the US, representing an important development in demography. The model was followed by several others models which were developed over the years [Gompertz, 1825; Makeham, 1860; Weibull, 1951; Beard, 1959; Vaupel et al, 1979, Kannisto, 1992]. The present study is trying to provide certain information over the best use of these types of models under specific hypothesis. We will focus on the methodology of estimating the parameters for Gompertz’ law of mortality and how well it can be fitted using data from Romania in 2012.