Causal Inference Between the Health Status and Living Arrangements of Elderly People in China

Abstract

This study investigates the causal relationship between health conditions and living arrangements of elderly people in contemporary China. The endogeneity between individuals’ living arrangements and health conditions and the dynamics of living arrangements for elderly people were considered using a cross-lagged panel design of the structural equation model with panel data from 2002, 2005, 2008, and 2011 waves of the Chinese Longitudinal Healthy Longevity Survey to explore the causality. During the analysis, to rule out the effect from the wave nonresponses among elderly people who are lost to follow-up and who are deceased, the missing at random (MAR) tests was firstly examined using multiple-group analysis by structural equation models. The full information maximum likelihood estimation under MAR conditions revealed that health status had a causal effect on people’s choice of living arrangements, instead of the other way around. Unhealthy conditions increased the probability of elderly people choosing to live with family members rather than living alone in China.