Date of Degree


Document Type


Degree Name





Janet Gornick

Subject Categories



climate change; climate change adaptation; income inequality; mapping vulnerability; spatial indicators; vulnerability


Social theorists suggest that income inequality within a society leads to a breakdown of social cohesion, spatial segregation, and as a result, uneven public resource access. I will assess whether this social phenomenon is important to consider when measuring vulnerability to climate change in urban, middle-income countries. To test this relationship, I create a flood hazard vulnerability index at the municipality level and determine whether income inequality, measured at the municipality level, is a predictor of municipality vulnerability to flood hazard. The flood hazard vulnerability index incorporates socioeconomic, built environment and natural environment data, providing a more holistic approach to vulnerability assessment. I draw on socioeconomic and spatial data from urban municipalities across 25 Brazilian states.

Using multi-level regression models, which account for state-level political economy impacts, as well as for the spatial dependence of flood hazard vulnerability, I test whether income inequality in a municipality, controlling for absolute poverty level and environmental hazards, predicts vulnerability to flooding, the most prevalent climate hazard in Brazil. I use several measures of income inequality to determine whether the effect of income inequality varies depending on where along the income distribution the income inequality lies. The measures of income inequality I select are: the Gini index, two measures of both bottom and top half income inequality and two specifications of the Atkinson index (inequality aversion parameter=0.5 and 1).

I find that the Gini index and the Atkinson index (inequality aversion parameter=0.5, calibrated to give more weight to the top end of the distribution), were the only two significant predictors of vulnerability. These results provide strong evidence in support of the two hypotheses in this dissertation, mainly that a certain type of income inequality is a predictor of vulnerability and that the location of the income inequality along the distribution does matter in terms of its impacts. It appears that top end, not bottom end income inequality significantly predicts vulnerability.

Next I dig further into the data and separately test each of the factors which comprise the composite vulnerability score: socioeconomic status, infrastructure quality and governance. This line of analysis yields some illuminating results. I find that while all types of income inequality positively predict socioeconomic status, when I control for absolute poverty, municipality size, and environmental conditions, top end and top half income inequality predict the poor governance component of the vulnerability index, the factor most closely correlated with the presence of slums, informal settlements and high population density.

In sum, these findings suggest that the level of absolute poverty does not fully explain the presence of slums, informal settlements, and high-population density within a municipality. Top end and top half income inequality also play a role. My data illustrate that considering income inequality, and specifically top end and half income inequality, as part of vulnerability assessments can significantly aid in crafting more effective, sustainable adaptation efforts by helping to better identify which municipalities are most vulnerable to climate change.

Included in

Sociology Commons