The Impact of Urban Sprawl on Disaster Relief Spending: An Exploratory Study

Abstract

Followers of urban affairs and public policy have written much over the years about the rise of suburbia and development beyond older city boundaries in the US, whether such development is called urban, suburban, or ex-urban sprawl. Many researchers have focused on various issues concerning sprawl, especially on the unintended consequences that new development has had for municipal finances, neighborhood income and residential segregation, and transportation planning, among other issues. Over the last decade, a new area in the literature on sprawl has focused on how the “built environment” of residential areas can impact health and emergency services. We contribute to this latest set of papers on sprawl by trying to empirically estimate the impact of sprawl in metropolitan regions on the Federal Emergency Management Agency’s (FEMA) spending on “public assistance.” This assistance encompasses spending on debris removal, emergency protective measures, and rehabilitating or rebuilding of infrastructure, public buildings, public utilities, parks and recreational areas, in post-disaster relief efforts. In our exploratory analysis, the results indicate that urban sprawl is a factor in influencing the level of FEMA’s public assistance spending.

Keywords::

• disasters
• disaster relief
• emergency management
• FEMA
• infrastructure
• urban sprawl

JEL Classification Codes::

• R14
• R28

Notes

¹ Counties and regions of U.S. territories (Puerto Rico, Virgin Islands, etc.) are also not included because they are not in the sprawl index list. We can make available all data analyzed for this paper upon request.

² Because on FEMA’s list there was only one earthquake that impacted just one county (Imperial County, California) and just one hurricane (Hurricane Alex) that impacted three counties in Texas, we dropped these events and their corresponding data from a third set of models to control for any confounding effects. However, the results did not change that much when we removed the variables for earthquake and hurricane.

³ See the list at www.fema.gov/disasters/grid/year/2010?field_disaster_type_term_tid_1=All.

⁴ Lambert, Srinivasan, and Katirai (2012) found that sprawled ex-urban areas were more prone to wildfires and delays in fire response.

⁵ FEMA states:

The Public Assistance Program provides grants to state and local governments and certain non-profit entities to assist them with the response to and recovery from disasters. Specifically, the program provides assistance for debris removal, emergency protective measures, and permanent restoration of infrastructure.

Eligible Applicants: Eligible applicants include state governments, local governments and any other political subdivision of the state, Native American tribes and Alaska Native Villages. Certain private non-profit (PNP) organizations may also receive assistance. Eligible PNPs include educational, utility, emergency, medical, temporary or permanent custodial care facilities (including those for the aged and disabled), irrigation, museums, zoos, community centers, libraries, homeless shelters, senior citizen centers, rehabilitation, shelter workshops and health and safety services and other PNP facilities that provide essential services of a governmental nature to the general public. PNPs that provide “critical services” (power, water — including water provided by an irrigation organization or facility, sewer, wastewater treatment, communications and emergency medical care) may apply directly to FEMA for a disaster grant. All other PNPs must first apply to the Small Business Administration (SBA) for a disaster loan. If the PNP is declined for a SBA loan or the loan does not cover all eligible damages, the applicant may reapply for FEMA assistance.

Public Assistance Process: As soon as practicable after the declaration, the state, assisted by FEMA, conducts the Applicant Briefings for state, local and PNP officials to inform them of the assistance available and how to apply for it. A Request for Public Assistance must be filed with the state within 30 days after the area is designated eligible for assistance. Following the Applicant’s Briefing, a Kickoff Meeting is conducted where damages will be discussed, needs assessed, and a plan of action put in place. A combined federal/state/local team proceeds with Project Formulation, which is the process of documenting the eligible facility, the eligible work, and the eligible cost for fixing the damages to every public or PNP facility identified by State or local representatives. The team prepares a Project Worksheet (PW) for each project.

Public Assistance Projects Categories:

• Category A: Debris removal

• Category B: Emergency protective measures

• Category C: Road Systems and Bridges

• Category D: Water control facilities

• Category E: Public buildings and contents

• Category F: Public utilities

• Category G: Parks, recreational, and other Items. (USFEMA 2014)

⁶ This is similar to a school-based letter grading scale, except that the latter starts with “F” as 0.

⁷ In disaster research, many types of infrastructure are called critical infrastructure — infrastructure that is vital to the functioning of a society, such as roads, bridges, public utilities, communications, etc. — and most of it is covered by FEMA’s public assistance expenditures/grants in the event of a disaster.

⁸ Unfortunately, for storms and flooding, the total depth or amount of flood water in inches in an area was not available.

⁹ After doing this for all of the interaction terms, and after looking at diagnostics for all variables in all of the regression models, we found no severe problem with multicollinearity with the variance inflation factors (VIFs) for each variable less than 5 (Studenmund 2005), and only one slightly above 5 but below 10, so as to satisfy a threshold of VIF<10 to indicate no signs of severe multicollinearity (Hair et al. 1995; Kennedy 1992; Marquardt 1970; Neter, Wasserman and Kutner 1989).

¹⁰ Proximity to water was not used by itself as independent variable because it was too highly correlated with the variable flooding.

¹¹ A CEI for each of the nine U.S. Climate Regions now exists, but there is only data beginning from 2013.

¹² We created two versions of a “severe weather index” that aggregated the different weather categories since many incidents involved more than just one type of severe weather. For example, snow storms — along with flooding or thunderstorms, severe winds, and flooding — impacted many communities. One weather index added together the values of relevant weather categories for an incident, so that if a county was affected by snow storms and flooding with each coded as 1, the categories of thunderstorms, hurricane, tornado, or earthquake would have 0 values. Another weather index aggregated the weather categories in our models by using factor analysis. Neither of these variables attained statistical significance, and so we used separate weather categories along with the interaction terms since some of these variables attained some level of statistical significance.

Additional information

Notes on contributors

Thomas E. Lambert

Thomas E. Lambert is an assistant professor of public administration at Northern Kentucky University, Highland Heights, KY. James Catchen is a dual-enrollment coordinator at Northern Kentucky University, and Victoria Vogelgesang is an alumnae relations director at Saint Ursula Academy in Cincinnati, OH. The authors would like to acknowledge the kind assistance of the Dean’s Office at the College of Arts and Sciences at Northern Kentucky University for providing generous financial assistance in the development of this paper.

James Catchen

Thomas E. Lambert is an assistant professor of public administration at Northern Kentucky University, Highland Heights, KY. James Catchen is a dual-enrollment coordinator at Northern Kentucky University, and Victoria Vogelgesang is an alumnae relations director at Saint Ursula Academy in Cincinnati, OH. The authors would like to acknowledge the kind assistance of the Dean’s Office at the College of Arts and Sciences at Northern Kentucky University for providing generous financial assistance in the development of this paper.

Victoria Vogelgesang

Thomas E. Lambert is an assistant professor of public administration at Northern Kentucky University, Highland Heights, KY. James Catchen is a dual-enrollment coordinator at Northern Kentucky University, and Victoria Vogelgesang is an alumnae relations director at Saint Ursula Academy in Cincinnati, OH. The authors would like to acknowledge the kind assistance of the Dean’s Office at the College of Arts and Sciences at Northern Kentucky University for providing generous financial assistance in the development of this paper.