Small, Slow, and Diminishing: The Effect of Democracy on the Under-Five Mortality Rate

Abstract

Theories of democratic accountability predict a positive effect of democracy on health outcomes. This prediction is at odds with published empirical estimates, which range from impossibly large to null effects, in part due to missing data, extreme counterfactuals, and lack of attention to dynamics. Here I use a new data set with substantially fewer missing observations, a research design robust to extreme counterfactuals, and a flexible dynamic specification. As expected I find democracy causes under-five mortality to fall. However, the estimated effect starts small, increases slowly, peaks after a generation, and diminishes thereafter. For example, had Morocco made a sustained transition to democracy in 1980, by 2003 its under-five mortality would have been some 10 percent lower than the actual realized outcome, saving 3,500 lives per year. This stylized fact calls for more nuanced theories linking democracy to health outcomes.

Keywords: democracy, under-five mortality, time-series-cross-section, stylized facts, missing data, extreme counterfactuals, matching, causality

JEL Classification: O10, O57, I00, J00, H40