Determinants of Child Health in Pakistan through Latent Variable Model Approach

This paper includes the estimation of demand production function for health in the case of Pakistan. We employ Multiple Indicator Multiple Cause (MIMIC) models which belong to the family of Structural Equation Models (SEM). The issue of unobservability of health has been addressed and it is found that latent variable with SEM is good representation of health state. Factor analysis has been used to extract measures of socio-economic status of parents due to the un-availability of income data. Various factors have been modeled in MIMIC which determines child health in Pakistan, also, their relationships with child’s transitory health with regard to their significance. Parents’ education, housing and type of residence are found out to be significant determinants. The contribution of parents’ education and housing towards child’s transitory health is positive, and children living in urban areas have better transitory health probably due to more access to better health inputs and infrastructure.