Comparison of estimators for probability of death used in actuarial science

In this paper a brief summary of the estimators for probability of death is given. Three estimators will be investigated, two of them are parametric ones. To derive a parametric estimator a distribution has to be assumed. Unfortunately the distributions occurring in real life differ from these assumed ones. It will be investigated how can these estimators be applied if we assume other distributions. The bias and efficiency of the estimators are analysed by using Monte Carlo simulation. The maximum likelihood estimator is the most common in the actuarial practice due to its appealing point estimation properties. The Kaplan–Meier estimation is a better choice, if the purpose is to give a better confidence interval.