Abstract:
The objective of this research is to consider whether there is difference for the mortality rates of Thai population from different birth cohorts. The process of research involves choosing the most appropriate model for Thai mortality rates from 5 candidate models consisting of Lee-Carter model, Renshew and Harberman model, Age-Period-Cohort model, Cairns-Blake-Dowd model and generalized Cairns-Blake-Dowd model. The data used in this study are the number of population and the number of death classified by age and sex in year 1963-2014. The value of Mean Absolute Percent Error (MAPE) and Bayesian Information Criterion (BIC) of each model are used to select the most appropriate model. After choosing the appropriate model, this research uses Autoregressive Integrated Moving Average model (ARIMA) to forecast the value of estimated parameters of the model and use them to forecast Thai mortality rates for the next 70 years. Lastly the forecasted mortality rates of different birth cohorts are used to calculate the premium for the example annuity insurance products. Then the value of premiums are compared to conclude how changing in mortality of different birth cohorts affects the value of premiums. The results revealed that RH model is appropriate for estimating Thai mortality rate. The forecasted mortality rates from RH model show that birth cohorts mortality rates tend to decrease which decreasing of female mortality rates is faster than males. Mortality rates decreasing by cohorts is slower than mortality rates decreasing by calendar years. The premiums calculated by forecasted mortality rates of later birth cohorts tend to increase because of decreasing in mortality rates by birth cohorts.