Determinants of Life Expectancy and Clustering: K-Means Panel for the Pandemic-Post-Covid-19 Pandemic Period in Indonesia
TL;DR
The results show that per capita expenditure, mean years of schooling, and proper sanitation positively affect life expectancy, while the number of health care facilities is insignificant, suggesting that service quality and equitable distribution may matter more than facility quantity.
Abstract
Inequalities in life expectancy across Indonesia remain a challenge for health development, particularly during and after the COVID-19 pandemic. Previous studies have examined socioeconomic, educational, sanitation, and healthcare factors, but findings remain inconsistent, with limited integration of determinant analysis and provincial clustering. This study examines the determinants of life expectancy and classifies Indonesian provinces based on similar health characteristics during 2020–2025. A quantitative approach was employed using panel data from 34 provinces. Panel regression was estimated using Common Effect, Fixed Effect, and Random Effect models, with model selection based on the Chow, Hausman, and Lagrange Multiplier tests, followed by Feasible Generalized Least Squares (FGLS). K-Means clustering was then applied to group provinces based on socioeconomic and health characteristics. The results show that per capita expenditure, mean years of schooling, and proper sanitation positively affect life expectancy, while the number of health care facilities is insignificant, suggesting that service quality and equitable distribution may matter more than facility quantity. K-Means identified three provincial clusters, revealing persistent disparities, particularly between eastern and more developed regions. The study contributes an integrated regression-clustering framework for targeted, regionally differentiated public health policies.