A Literature Review of Disaster-Induced Business Interruption and an Exploratory Analysis of the Effects of the 2004 Atlantic Hurricane Season on Florida Establishments at the Zip Code Level

Abstract

Businesses play a vital role in community resilience. They provide supplies and services in the wake of a major disruption, employment to survivors, and are necessary for a community to begin a full economic recovery. At the same time, businesses themselves are subject to the effects of disasters and disruptive events. They may suffer physical damage to their place of business, lose employees as a result of death and/or loss of homes, and lose customers due to forced closures while repairs are made and/or lifeline infrastructures are brought back to operational status. These impacts generally serve to accelerate any preexisting trends in the region, and can linger for years. Thus understanding the factors that allow a business to survive is vital to establishing a resilient economy. season was a distinct and severe enough event that its impacts are distinguishable from general trends in the data. Analysis was also broken into coastal regions and inland regions to lessen any possible endogeneity. The change in the number of establishments was selected as the dependent variable because establishment level data on closures was unavailable.A difference in differences (DiD), as well as a graphical analysis on the change in the number of establishments, support the view that the 2004 hurricane season was a significant disruption. However due to the multitude of convoluting factors affecting a state's economy, it is impossible to attribute the full effect to the 2004 hurricane season. Fixed effect regressions using first differences were also run looking at the impact of demographic factors on the change in number of establishments. Regression results were rendered inconsistent by an endogeneity issue; however, the general method here did produce interesting results when incorporating year-specific effects: in many cases, these results agreed with the DiD results and graphical analysis. The regression results imply demographics have a mixed impact, bearing in mind the estimator was inconsistent due to endogeneity. Of more value is that the derived method is general enough that, if a complete data set were present, it could produce meaningful results.