When do democratic transitions reduce or increase child mortality? Exploring the role of non-violent resistance.

What explains variation across countries in the effect of democratization on child mortality rates? Democratic transitions, on average, improve health outcomes but there is substantial variation across countries in whether democratization leads to lower-than-expected child mortality post-transition. As yet, there is no convincing quantitative explanation for this variation. In this paper, we argue that whether you have a protest-led or violence-led democratic transition alters the trajectory of child mortality post-transition. Our paper makes two contributions. First, we offer a more detailed account of how the type of resistance movement promoting regime change affects health post-transition. We also draw on novel data to categorise the movements producing democratic transitions as violent or peaceful, moving beyond earlier work which operationalised peaceful democratizations in terms of battle-related deaths. Second, we extend earlier research by examining whether the nature of the democratization movement constitutes a necessary cause of higher or lower-than-expected child mortality following democratization. Across 51 transitions, countries that have a protest-led transition have lower-than-expected child mortality rates after the transition to democracy than countries with other kinds of movements (β = -0.17, p = 0.003). Countries with violence-led transitions, meanwhile, have, on average, higher-than-expected child mortality rates after their transition (β = 0.20, p = 0.001). These associations hold when we adjust for covariates (including all possible combinations of various confounding variables). We also find evidence that protest-led transitions may be a necessary condition for avoiding increased child mortality post-transition. Finally, we conduct a deviant case analysis of transitions that appear to be contrary to our theory, finding that these cases are likely instances of measurement error. Democratization may not always improve health, but such health improvements are more likely when regime change is protest-led. This is because such movements are more likely to build broad coalitions committed to consensual politics post-transition, a critical feature of successful democracies.