Where Water Meets The Wound: Mapping the Political Geography of Disaster Aid in the Gulf Coast
- Dominique Williams
- Aug 15
- 9 min read
I spent 8 weeks over the summer conducting independent research in the Mobilisation and Political Economy Summer REU at the University of Pittsburgh. I believe that all research, even inconclusive and null findings, has a role to play in advancing knowledge. Therefore I have decided to condense my key findings into this digestible summary with the hopes that it can be useful to others.
Table of Contents
INTRODUCTION
Climate change is intensifying, with major disasters like wildfires, hurricanes, and floods becoming more frequent. Beyond these visible events, climate change also exacerbates inequalities through droughts, water scarcity, food insecurity, and rising prices from supply shocks. Natural hazards often worsen pre-existing vulnerabilities that are shaped by government policies and post-disaster responses.
In the United States, especially the US Gulf Coast region, the impacts of hurricanes are segregated by race, class, and place-based hazards such as degraded environments, proximity to petrochemical and other exploitative industries, and poor urban planning. Hurricane Harvey was among the most devastating storms to strike the Gulf Coast, causing on of the region’s largest flood event. Studies showed that racial and ethnic minority populations exhibited the lowest levels of disaster preparedness before the storm and experienced greater challenges to recovery.
The question motivating this research proposal asks whether differences in voter turnout across socially vulnerable populations in Texas and Louisiana affect the allocation of aid in the aftermath of Hurricane Harvey. This research study builds on existing literature where voter turnout is linked to government responsiveness in socially vulnerable communities.
BACKGROUND AND LITERATURE REVIEW
To address the research question proposed in this project, this study draws on three topics of theory: (1) government responsiveness, (2) social, environmental and climate vulnerability and lastly (3) racial capitalism.
1. Government Responsiveness
This is the field of study concerned with how incumbent politicians respond to the needs of their constituents especially in relation to addressing disasters. Research in this field indicates that elected officials tend to exhibit higher levels of sensitivity to areas with perceived greater political benefits. This is because politicians understand that voters punish incumbents who they believe did not act quickly or strongly enough to protect them. These voters also reward the politicians who they deemed were efficient at handling the crisis.
The gap that this research hopes to address is by mapping the ways in which government responsiveness is inadequate in addressing the needs of black, hispanic, low-income and high climate-risk populations broadly referred to in this research as socially vulnerable populations. Existing research shows that racial minority and low-income communities experience greater impacts of disaster, but elicit lower levels government responsiveness after disasters in the form of disaster recovery spending.
2. Social Vulnerability, Environmental Justice and Climate Justice
Modern disaster theorists believe that existing risk factors such as poverty, disability, and climate risk, before a natural disaster are exacerbated by the negative impacts caused by natural hazards in the aftermath. This can occur when climate risk interacts with social vulnerability of these populations by limiting their access to resources and reducing their ability to advocate in governance structures.
Access to resources includes all tangible, intangible, private, and public goods such as emergency savings, emergency evacuation plans, flood and property insurance, emergency food supply, etc. As basic as these resources may seem, preparation against climate threats are more difficult for low-income households because they may lack the required income, time, language abilities, knowledge, etc.
In addition to economic means, one's social power can also affect how they cope with disasters. Social and economic power are nearly indistinguishable since social power – the ability to influence or coerce different groups – is derived from and reproduces the social hierarchies that create inequalities in economic power.
Depending on the power one possess' they can influence the processes that perpetuate marginalisation and climate risk, such as adaptation planning and access to funding to cope with, and respond to climate-related impacts. Therefore, government responsiveness can create vulnerability, especially if electoral representation and political marginalisation negatively impacts a community's climate resiliency.
3. Racial Capitalism
This is a political economy framework that focuses on the interplay of politics and economics. As a political economy framework, racial capitalism helps understand how one's race can influence the quality of environment they live in, the harm they experience due to climate change, their exposure to industrial pollution, the amount of poverty faced, and in the case of this research: whether or not their local government promptly responds to their needs in the aftermath of a hurricane.
The entire economic, environmental, and urban landscape of the Gulf Coast is oriented around maximising the economic output of petrochemical industries through lax laws and the absence of strong environmental or social protections. This leads to the residents of this region feeling abandoned and neglected by their local governments who often have close ties to the polluting industries.
Therefore, this research proposal incorporates existing scholars' understanding of place-based exclusive policies to form a political geography that physically defines who experiences protection from the government and who does not. Groups that face class and race-based vulnerabilities may express their frustration with the political system by refusing to participate in politics that they believe fail to protect them or their interests.
HYPOTHESIS
This research project seeks to show how political behaviours within marginalised populations affect the uneven distribution of aid, thereby reinforcing inequality in the aftermath of crises. By focusing on the Gulf Coast, where vulnerable populations are also exposed to heightened levels of industrial pollution, this research also hopes to shed light on how overlapping geographies of social, environmental, and climate risk relate to political engagement and mobilization.
My hypothesis is that counties with higher levels of voter turnout will receive higher levels of disaster recovery aid and that this relationship between voter turnout and aid distribution will be more pronounced in communities with higher levels of social vulnerability.
RESEARCH METHODOLOGY
The bottom portion of the picture above illuminates the four main sources of data I collected when creating the dataset for this study. These databases are
The Arizona State University’s Spatial Hazard Events and Loss Database for the United States (SHELDUS) which included data about hurricane damage.
The 2016 CDC ASTDR's Social Vulnerability Database which included data on various social risk indicators.
The State Department websites of Louisiana and Texas where I collected data on voter registration and voter turnout.
The Federal Emergency Management Agency (FEMA) Hazard Mitigation Grant Program (HMGP) dataset which included data on the amount of aid distributed through this program.
This data was collected for 73 counties/parishes which were selected based on the criteria that they qualified for Public Assistance from FEMA when the federal government made the disaster declaration for Hurricane Harvey in Texas and Tropical Storm Harvey in Louisiana.
The independent variables are 1) voter turnout in 2016 elections before Hurricane Harvey in 2017 and 2) the social vulnerability of each county. Damage and population were included as control variables.
KEY FINDINGS AND ANALYSIS
Bivariate Analysis
A bivariate analysis looks at the relationship between two variables. Scatterplots were used to easily observe patterns in the larger data sample.
These three graphs would suggest that the sample behaves in ways that existing studies would predict. Fig. 1 shows that aid and voter turnout have a positive relationship meaning that places with higher voter turnout received more aid. Fig. 2 shows that places that are more vulnerable received less aid and Fig. 3 shows that places with high levels of vulnerability had lower levels of voter turnout.
Multivariate Regression
The regression summary below shows the relationships between the independent and control variables and the dependent variable Federal Aid. According to this model, social vulnerability has a negative relationship to federal aid as the hypothesis would have predicted.
However this model also suggests patterns opposite to the hypothesis'. In this model voter turnout has a negative relationship instead of a positive one with Federal Aid. Meaning that within the cases studied, counties with higher levels of voter turnout received less recovery aid.
However it is important to note that the p-values of voter turnout and social vulnerability are greater that o.5 which means that they are not statistically significant.
A secondary regression conducted gave additional insight into the interplay of these variables. An interaction term looks at two variables in relation to each other and their combined influence on the dependent variable. In the regression below social vulnerability and voter turnout has a positive relationship. Therefore although voter turnout on a whole has a negative relationship with federal aid, this new regression suggests that within socially vulnerable communities, higher voter turnout can lead to higher levels of disaster aid relief spending.
Interpretation of Findings
Vulnerability and Disaster Response
In the counties observed, the communities that had more pre-existing inequalities before Hurricane Harvey were less likely to receive government aid compared to counties with lower levels of vulnerability.
A supplementary scatterplot showed a negative relationship between damage and vulnerability, potentially because property damage is one of the main ways that disaster destruction is measured and property values are lower in low-income communities. The model shows a strong, significant relationship between damage and federal aid allocation, suggesting that governments are more responsive to areas with high levels of reported damage.
Vulnerable communities might be at risk of being overlooked because of their lower damage amounts, and because their harms often manifest in non-monetary ways, such as food insecurity, contaminants from chemical plant releases, disruptions in public services, and public infrastructure damage.
Population Size and Disaster Response
The regression analysis reveals a strong correlation between population size and government aid, with a p-value less than 0.005. This suggests that governments are more responsive to larger populations, as they represent a bigger voting base. It can be inferred that county officials are more inclined to respond to more urban areas, while rural and sparsely populated counties and communities are less prioritised.
Voter Turnout and Disaster Response
The hypothesis predicted that counties with higher levels of voter turnout would receive more government disaster recovery aid. While this was suggested in the bivariate analysis, the multivariate regression showed a negative relationship between the two. These conflicting results suggest a hidden underlying mechanism between the communities that are more likely to vote and those that require more recovery aid.
Adopting a Racialised Policy Feedback Framework (RPFF) can assist the understanding of the interactions between voter turnout, government responsiveness, and social vulnerability. A policy feedback loop is the process by which existing laws impact the formation of future laws. In a racialised policy feedback framework, existing policies limit the ability of black communities and other marginalised populations to voice their dissent for new and existing policies, thus becoming more marginalised by future policies that may negatively impact them.
Texas and Louisiana exhibit high levels of inequalities perpetuated by lingering Jim Crow legislation, voter suppression, redlining, and the placement of harmful industries. These laws concentrate hazards along class, race and place. Social vulnerability indicators such as race, income, and exposure to pollution might suppress people's behavioural feedback by eroding trust between communities and the government, depressing their willingness to engage in political action.
RPFF could explain why voter turnout was no longer positively related to federal aid in a multivariate analysis, since voter turnout might only be a small factor in a much larger mechanism of political engagement, mobilisation, political distrust, racial and spatial risk factors and class. This is the gap that this research and future research in this area will contribute to.
Applications of this Research:
Continued research on the policy feedback systems within these communities could unlock new knowledge on the barriers that exist between these communities and the assertion of their policy preferences. Removing these barriers becomes increasingly important as the risks to these communities intensify due to climate change.
Community leaders should invest in mutual aid networks that can serve populations rendered invisible to government institutions. In the aftermath of Hurricane Maria, where disaster relief efforts were insufficient, local organisations, churches and private citizens ran food pantries and other services to aid those in need.
Addressing the root of the problem, local activists and organisers should evaluate the benefits and disadvantages of electoral and non-electoral modes of signalling policy preferences. As climate change risk steadily increases, governments must be careful to mitigate rather than exacerbate the additional risk it poses to existing vulnerable communities.
BIBLIOGRAPHY
Please utilise this link to my bibliography to access my sources:
AKNOWLEDGEMENTS
Thank you to my research proposal advisor Professor Fernando Tormos-Aponte for your guidance in formulating and pursuing this research. Additional thanks to Mauricio Zavaleta and Professor Nick Rogers for their assistance in the applying quantitative skills to this research methodology. My heartfelt gratitude to the TA's, coordinators and assistants who supported this year's Mobilisation and Political Economy REU Program. Finally, thank you to faculty, mentors, my fellow cohort and all others who supported me in this journey.
Comments