Pemodelan Angka Harapan Hidup di Provinsi Papua dengan Pendekatan Model Data Panel Dinamis Spasial

Abstract

Life expectancy rate is one of the indicators used to comprehensively measure the health status of the community in an area. Life expectancy is used as a reference in planning health programs and as an evaluation of the government's performance in improving the welfare of the population in general and improving health status in particular. One of the provinces with the lowest life expectancy achievement is Papua Province which is in the second most down position nationally from 2014 to 2019. The achievement of life expectancy in Papua province varies widely with different characteristics between regions. This achievement is very different from other provincial factors, so it is necessary to identify the contributing factors. In this research, life expectancy will be modelled using spatial regression analysis of dynamic panel data. Panel data regression is used for the relationship between two or more in the data panel. In contrast, spatial regression is used for the relationship between several things taking into account the spatial effect. In other words, this research was conducted to examine the relationship between several changes in the data panel by considering the spatial aspect. The data used in this study is secondary data from the Indonesian Central Statistics Agency for 2012-2019. The parameter estimation method used is Maximum Likelihood (ML). The variables that are thought to affect life expectancy in 29 districts/cities in Papua Province are the percentage of the sick population, the percentage of households that use a proper drinking water source, school year expectations, average years of schooling, and population density. The results showed a spatial dependence on the response and explanatory variables. The modeling of the life expectancy of the Papua province in 2012-2019 is the Dynamic Spatial Durbin Model (MDSD) with a weighting matrix for the two closest neighbors. The modeling results show that the estimation of the effect of spatial dependence and time lag, namely ρ and λ has a significant impact. The variables that affect life expectancy are expected years of schooling, population density, and the percentage of households with proper sanitation. This modeling can explain the diversity of the model by 73.57%. The marginal effect shows that the variable with the most significant influence on life expectancy is school year expectancy. Further research can add other variables that can affect life expectancy in terms of socio-economic, environmental, and others and can build models that are robust to outliers.