2025 Theses Doctoral

Essays on environmental change and social inequality

Davis, William Matthew Alampay

This dissertation is comprised of three essays each of which may be described as empirical investigations relating environmental change to a distinct concept of inequality: economic inequality as measured by the global distribution of income; political inequality as reflected in the demand for and quality of democratic institutions; and patriarchal cultural inequality as expressed through domestic violence.

Chapter 1 explores the relationship between anthropogenic climate change and global inequality, a subject bridging two of the defining challenges of the 21st century but that nonetheless remains virtually unstudied at subnational levels. To make overdue progress addressing this research gap, I adapt recently developed methods for estimating climate impacts to overcome long-standing empirical limitations and apply these refinements to fine-scaled data on globally representative income distributions.

I document new evidence that temperature shocks significantly and persistently impact distributions of income within countries, an effect driven by concentrations of harm onto the lowest income-earners in warm climates as well as a surprising vulnerability of the top 1% in these countries to environmental shocks. Integrating these inequality effects over observed distributions of income and the spatial incidence of global warming between 1980 and 2016, I find that climate change has regressively redistributed global shares in a reduced stock of income both between and within countries largely by depriving the world's poorest of economic opportunity that would otherwise have been available. For the poorest percentile of positive income-earners in our data, we find that income levels absent asymmetrical warming since 1980 would be 29% [18, 41] higher than realized. Through an inequality decomposition exercise, I estimate that climate change over this period has reduced between-country inequality by 8.7% [4.9, 13.3] and within-country inequality by 2.6% [0.0, 5.6]; to the best of knowledge, this decomposition and especially this latter quantity have not previously been estimated. Altogether, these results constitute the most comprehensive evidence yet of the regressive impact of climate change.

Despite widespread regard of climate change as a threat to global stability, the nature of the relationship between environmental change, political preferences, and regime change is undertheorized in environmental political economy. In Chapter 2, I consider two possible frameworks by repurposing influential existing theories of regime change. Under a "mercury uprising" hypothesis, adverse environmental shocks diminish the opportunity cost of contesting autocratic rule thereby improving prospects for democratic reform. On the other hand, a "state of exception" hypothesis would suggest that the public responds to these shocks with a willingness to compromise civil liberties for authoritarian features of governance which may more credibly ensure security thereby leading to a diminution of democratic institutions.

I empirically test the relative merits of these theories first by measuring the impacts of identified temperature shocks on survey-based proxies for the demand for democracy. Next, I use multi-valued and dichotomous measures of democracy to investigate the dynamic effects of these shocks on the democratic quality of institutions. The results from these analyses consistently support the latter theory. In particular, I find that a positive 0.5°C temperature shock increases the propensity for autocracies in temperate geographies to democratize by between 1.5-3.7 percentage points depending on the operating definition of democracy but reduces this propensity for their tropical counterparts by 1.6-2.4 percentage points. The same shocks also increase the propensity for warm-climate democracies to break down by 0.6-2.0 percentage points. In these cases, effects are substantial and significant but exhibit very limited persistence beyond the year of the shock after accounting for autocorrelation in the shock itself. Finally, we find that the same shocks have null effects on the propensity for temperate democracies to undergo reversal, suggesting these institutions have historically proven impervious to environmental disruption.

In Chapter 3, I explore how explore how drought-induced weather shocks impact spousal violence in a previously never-violent relationship. Here, I combine high-resolution weather data with comprehensive duration data on the incidence of physical and emotional violence in India to estimate how drought-induced negative income shocks may affect these health outcomes.

My findings suggest that negative income shocks substantially reduce the hazard into marital violence. Specifically, a negative precipitation shock of a magnitude one might ordinarily expect once every 4-5 years on average reduces the risk of a never-violent relationship becoming violent by 0.037 percentage points, representing a 70% reduction in the baseline hazard of this transition during the formative years of a marriage.

I also replicate a recently published study relating these weather shocks to the risk of child marriage using updated methods and much-improved data. I find that these effects are milder but more persistent than characterized in the original study. In addition, I find that this result is reproduced in Sub-Saharan African countries, contradicting the headline result of the original study which had reported opposite-signed effects in the two settings and attributed the difference to opposite consumption smoothing incentives motivated by opposite consumption-smoothing incentives imposed by the cultural practices of dowry and bride price payments. Instead, secondary results suggest that child marriage seems to be even less economically motivated than marriages occurring later in life.