Explaining the Decline of Child Mortality in 44 Developing Countries: A Bayesian Extension of Oaxaca Decomposition Methods

TL;DR

This study uses a Bayesian extension of Oaxaca decomposition to analyze the decline in infant mortality in 42 LMICs, revealing that improved health interventions and technical progress, rather than changes in population characteristics, primarily drove the decline.

Contribution

It introduces a novel Bayesian Oaxaca decomposition method to identify the main drivers of infant mortality decline in developing countries.

Findings

Decline mainly due to reduced infant death propensity among parents with similar characteristics.

Technical progress and public health interventions are key contributors.

Changes in population characteristics had minimal impact.

Abstract

We investigate the decline of infant mortality in 42 low and middle income countries (LMIC) using detailed micro data from 84 Demographic and Health Surveys. We estimate infant mortality risk for each infant in our data and develop a novel extension of Oaxaca decomposition to understand the sources of these changes. We find that the decline in infant mortality is due to a declining propensity for parents with given characteristics to experience the death of an infant rather than due to changes in the distributions of these characteristics over time. Our results suggest that technical progress and policy health interventions in the form of public goods are the main drivers of the the recent decline in infant mortality in LMIC.