Natural Disasters, Bankruptcy, and Disaster Aid: Evidence from the United States

Natural disasters are economic shocks that destroy physical assets - over $250 billion in property damage between 2000 and 2009 (2014$). Nearly 1% of all American households file for bankruptcy each year, often as a result of "adverse events" - shocks to household income as small as $500. Two previous studies suggest a correlation between major natural disasters and bankruptcy, but this study is the first to measure the effect of natural disasters on bankruptcy using national data over a significant time period. One means of aiding recovery after natural disasters, federal disaster aid, has increased dramatically but there has been little research into the effectiveness of this aid. What has been shown is that presidential disaster declarations are politically motivated, although data limitations have prohibited similar studies of the motivations behind Small Business Administration (SBA) disaster declarations, another major component of federal aid. In this dissertation I construct a panel of disasters, disaster aid, and bankruptcy to show that large natural disasters (causing over $50 per capita in damage to a judicial district) result in an immediate 4% increase in bankruptcy. Importantly, I take account of both variation in district baseline levels and district-specific trends in bankruptcy. Two years after the damage, however, bankruptcy is lower than the trend. Aid in the form of both grants and loans does not reduce bankruptcy; in fact these measures are correlated with an increase in bankruptcy. This is because aid declarations are better measures of concentrated damage than my per capita damage variable, which is measured at the (very large) judicial district level. I collect new data on SBA disaster declarations from the Federal Register and via a Freedom of Information Act request, which I then use to measure the political motivations behind SBA declarations. I show that they are susceptible to the same political motivations as presidential declarations. Disasters occurring in reelection years, under Democratic administrations, and in swing states are more likely to receive a declaration than other disasters. These political influences are much weaker after the passage of the 1988 Stafford Act, after which presidential disaster declarations begin to respond more to political incentives. This suggests that SBA and presidential declarations are substitutes in achieving political goals.