Proyeksi Tingkat Kematian di Indonesia Menggunakan Metode Holt-Winters Smoothing Exponential dan Moving Average

Abstract

Inaccurate predictions would cause the insurance companies to incur huge losses and may lead to expensive premiums for which low-income consumers are unable to insure themselves. The ability to predict mortality rates accurately allows the insurance companies to take preventive steps to introduce new policies with reasonable prices. It is hoped that by carrying out mortality projections, losses caused by longevity risk in the life insurance industry would be minimized. This study used secondary data obtained from the World Health Organization (WHO) website in the Mortality and Global Health Estimates category with the sub-topic Life Table by Country Indonesia. In this paper, several models are used to predict the mortality rate in a case study population in Indonesia, namely the Moving Average and Exponential Smoothing forecasting methods. The results obtained are the best method for predicting mortality rates is by using the Exponential Smoothing method with the MAPE value of Exponential Smoothing is smaller than the MAPE value on the Moving Average. The results of this mortality projection will later be used to obtain the distribution of life expectations and the premium price of life annuities.