Essays on Natural Disasters and Fiscal Resilience

Abstract

This dissertation attempts to bridge two fields within public policy, sustainability and public budgeting and finance. Natural disasters have become more frequent, intense, and costly over the last few decades, and this study contributes to the discourse through a public finance lens. I develop three separate but interrelated essays, which seek to understand how environmental shocks like natural disasters affect government budgets, and what actions governments should take to maintain fiscal resilience and sustainability. The first chapter evaluates the impact of flooding events on state government fiscal conditions. I construct a dataset with panel data from 50 U.S. states between 1997 to 2020 and employ a two-way fixed-effects model. The results show that increased severity of flooding leads to higher intergovernmental revenue two years after the disaster, but states do not necessarily pass down more funds to localities. There is also evidence that an increased flooding severity coincides with increased state tax revenue during and one year after the disaster. While federal assistance can help states stabilize during and after an emergency, I argue that it may also create problematic incentives. Specifically, federal transfers may discourage states from allocating sufficient funds for ex-ante flood mitigation. The second chapter examines whether and how government attention to hurricanes developed over time. I analyze budget documents from 2005 to 2020 across seven hurricane-prone states in the U.S.: Alabama, Florida, Georgia, Louisiana, North Carolina, South Carolina, and Texas. While prior studies focus on budget data, my study is the first to exploit the narrative portions of the budget documents using textual analysis. Computational text analysis is used to identify and quantify government attention toward different policy areas, including disaster management in connection with hurricanes. Cases from Louisiana, Alabama, and Florida show that an increased incidence of major hurricanes coincides with higher attention to hurricanes in the budget documents. There is also evidence that the attention to hurricanes is associated more with reactive rather than proactive measures, calling into question the general preparedness of most state governments for future disasters. The third chapter surveys the government’s disclosure of risks associated with natural disasters. It is intended as an exploratory analysis using audited financial reports from 50 U.S. states between 2002 and 2016. Audited financial reports are more technical and less political than the budget documents examined in my second chapter. I combine qualitative analysis with computational text analysis to understand the narrative portions of the source document. The results show that state governments do not widely disclose information regarding disaster risk despite the increasing threat of natural disasters. States that disclosed disaster risk in the past tend to disclose the same information in the future and past years, suggesting significant inertia in financial reporting. Finally, based on selected cases in hurricane-prone states, the study finds proactive as well as reactive risk disclosure, suggesting the need for governmental accounting standards to incorporate mechanisms that specify how disaster risk should be recognized, measured, and presented in the financial reports.