Document Type : Original Research Paper

Authors

• A. Komijani
• M. Koosheshi
• L. Niakan

Department of Economic Sciences, University of Tehran, Tehran, Iran

Abstract

Mortality forecasts are made in two ways: one is indirect forecasting through life expectancy forecasting and then converting it to age-specific death rate, and the second is direct forecasting of mortality rates. In the first approach, it is usually assumed that in the post-transition period (which is usually a level higher than 70 years for life expectancy at birth), the increase in life expectancy slows down, and as it approaches the exponential, its increase will be insignificant. In the second approach, the age-specific death rate is predicted and the life expectancy is obtained using the direct method of building the life table. As a rule, the logical and correct method of estimating or predicting life expectancy is the second approach. In addition, the indirect method has difficulties and generally larger errors. The purpose of this article is to use the second approach, and in this direction, using Iran's mortality data, the Lee-Carter model is used to predict the mortality rate. This model is based on two main elements: one is time and the other is age. The strict and important assumption of the model is the relative stability of the age-of-death pattern, and various evaluations have shown that in these conditions, the error of the Lee-Carter model is lower than any other method for directly predicting the mortality rate. In the case of Iran, the estimates show that the estimation error is insignificant, although this element is higher in some ages. Testing the hypothesis of the stability of Iran's mortality age pattern is an assumption that can be the subject of further research.

Keywords

• Life expectancy
• age specific mortality rate
• Lee-Carter model
• mortality prediction methods