Examination of Korean Mortality Forecasting Models

The purposes of this study are to examine a method to overcome the shortage of historical data on mortality of the elderly and to find the best model to forecast Korean mortality rates overall. To extend the mortality for ages 75 and over, we test two methods of estimating death probabilities: the 2-parameter logistic model and the Brass-Logit model. Based on the Mean Absolute Percent Error (MAPE), the logistic model has better performance than the Brass-Logit model. Four stochastic forecasting models (the Lee-Carter Model, the adjusted Lee-Carter Model, the Lee-Miller Model, and the Coherent Lee-Carter Model) are fitted to the period 1970-2010. The forecasts are compared to actual mortality for that period. The results of this evaluation show that the Coherent Lee-Carter Model is consistently more accurate in forecasting Korean mortality rates than other compared models. The Coherent Lee-Carter model yields a higher life expectancy at birth for both sexes and a larger difference between sexes than other models in which sex differentials diminish rapidly.