Modeling Regional Patterns of Post-Disaster Permanent Housing Reconstruction Outcomes in the U.S.

Srijesh Pradhan, Erin Arneson, Ph.D., and Rodolfo Valdes-Vasquez, Ph.D.
Colorado State University
Fort Collins, Colorado

Availability of construction and capital resources are fundamental to carry out post-disaster residential repair or reconstruction tasks in a market-driven reconstruction environment such as the U.S. The capacity of regional construction markets determines the availability of construction resources (labor and material) to meet the reconstruction demand. In contrast, the socioeconomic conditions of households assess the availability of capital resources. Limited quantitative studies are examining the influence of pre-disaster construction and capital resource availability and post-disaster residential reconstruction outcomes at the regional scale in the U.S. Since the construction capacity and socioeconomic conditions of households vary geographically across the U.S., the relationship between pre-disaster resource availability and reconstruction outcomes can also vary across the disaster-affected regions. However, very few studies have explored the geographical variation of the relationships between pre-disaster resource availability and residential reconstruction outcomes.

The objectives of this study are to: (1) establish global relationships between pre-disaster construction and capital resource availability and post-disaster permanent housing reconstruction outcomes using Ordinary Least Squares (OLS) regression model, and (2) determine if the relationships between pre-disaster resource availability and post-disaster permanent housing reconstruction outcomes vary across the study region using Geographically Weighted Regression (GWR) model.

The Northeast census region of the U.S., comprising of eight disaster-affected states, was used as a case study region as it was hit by multiple billion-dollar disasters between 2011-2012 (e.g., Hurricane Irene and Hurricane Sandy). Counties were used as a geographical unit of analysis for this study. County-level availability of construction resources was measured for the pre-disaster baseline year 2010 using economic indices of the regional construction industry (e.g., Location Quotient of construction employment, construction establishments, and wholesale establishments). County-level availability of capital resources was measured for the pre-disaster year 2010 using socioeconomic indicators (e.g., median household income, education status of households, and home mortgage status). County-level post-disaster permanent housing reconstruction outcomes were measured as the change in median home values from the pre-disaster year 2010 to the post-disaster year 2013 using a two-year reconstruction timeframe. A global OLS regression model was created using construction and socioeconomic resourcing variables as predictor variables and reconstruction outcomes as the outcome variable. A local GWR model was constructed using the same variables to explore the geographical variation in the relationships between predictor and outcome variables across the study region counties.

The OLS model revealed that pre-disaster construction and capital resource availability (measured through construction industry and socioeconomic resourcing variables) significantly influenced post-disaster reconstruction outcomes. The OLS model explained about 58% of the variation in median home growth rates values through construction and socioeconomic resourcing variables. The GWR model, with an adjusted R-squared value of 0.79, revealed that the relationships between resource availability and reconstruction outcomes varied across the study region counties.

This study contributes to the theoretical knowledge by exploring spatial heterogeneity in the relationships between resource availability and reconstruction outcomes. Exploration of such spatially varying relationships can help inform local policy and aid city planners, homeowners, and reconstruction stakeholders in the decision-making process to enhance disaster resilience of communities.

Keywords: Residential housing reconstruction, Disasters, GIS