Date of Conferral





Public Policy and Administration


George Larkin


Natural disasters expose the fact that poverty, race, gender, and other indicators of social disadvantage are linked to the population of citizens who struggle the most to recover after a disaster, yet these factors are not accounted for in public policy that guides decision making related to federal assistance to residents affected by a disaster. This study used neural networks as a research strategy to determine whether the current policies under the Stafford Act related to assistance comply with Congressional intent and law that uses a formula for assistance distribution, and whether human factors such as culture, measured as residing in a non-white zip code according to Census tract data, are considered in decision making regarding assistance. Data from FEMA related to the recovery from Hurricane Irene in 2011 were used as the basis for the model. The neural network analysis of this study indicated that federal assistance decisions after the Hurricane Irene event tended to focus on the adjusted property value and actual dollar value of losses as the determining factor in decisions. Focusing on the actual dollar value of losses is consistent with the formulaic approach codified in public law, but this approach overshadows important human factors such as living in a primarily non-white zip code and the availability of temporary housing. This study underscores the notion that the public policy works the way it is intended, but it fails to accommodate human and social factors. As a consequence, the existing policy is legally equitable, but it is not necessarily morally fair to those impacted by disasters. The positive social change implications of this study include recommendations to federal policy makers to more equitably structure recovery efforts in alignment with the human environment of communities rather than a primary focus on cost and value of real property.