Talk will cover:

• Motivation and Research Question
• Data
• Empirical Strategy
• Results
• Economic Relevance

• Climate change implies more frequent and more damaging disasters (Van Aalst 2006)
• Disasters do not discriminate with respect to wealth, but institutions and wealth helps mitigate deaths (Kahn 2005)
• Growing fiscal and social costs within the United States
  • More frequent FEMA disaster declarations
  • Higher federal costs for declared disasters
  • Non-disaster government transfers increases may understate true cost (Deryugina 2016)

• Some evidence that people move after big disasters
  • Boustan et al. (2017)
  • Carleton and Hsiang (2016)
• Rational actors respond to disasters by moving to new locations
  • Because their homes are destroyed?
  • Because the local economy is harmed?
  • Because they have updated their priors about risks?
• Move people out of dangerous areas, but does this lower future disaster costs?
• All else equal, higher exposure areas should see lower migration rates

• Disaster exposure is spatially correlated
• Disasters are negative income shocks and long moves are costly
• People tend to move relatively short distances
• If disasters induce shorter moves, migration may not lower future disaster costs
• Shorter moves may not lower costs
  • Key question: where do disaster migrants locate to?
  • And does migration mitigate impacts of disasters or simply rearrange the deck chairs?

• Category 3 Hurricane at landing (sustained winds of 100–140 miles per hour)
• Landed 25 August 2005 in Louisiana
• Total damage estimated at $108 billion (Knobb, Rhome, and Brown 2005)
• Deaths appear to be at least 1,500 (Beven et al. 2008)
• Triggered response of over $5.7 billion in FEMA disaster relief

• Data from IRS county-to-county migration used, tracks filed tax returns at FIPS code from year to year
• Typical outmigration from New Orleans area from 2000 to 2010 excluding 2005 averaged 21,482 outmigrants per year (average population of 1,421,693)
  • Parishes - Jefferson, Lafourche, Orleans, Plaquemines, St. Bernard, St. Charles, St. John the Baptist, and St. Tammany, and Terrebonne
  • Fairly stable spatial distribution - nearby and large metro regions
• In 2005, total of 181,854 outmigrants
  • Similar spatial pattern exists plus evident overflow

Beginings from Newtonian physics, which imply

\[ F = G \frac{m_1 m_2}{r^2} \]

Princeton Astronomer James Q. Stewart observed the composition of students in his classroom (J. Q. Stewart 1941):

\[ M_{ij} = k \frac{P_{i}^{\beta_1} P_{j}^{\beta_2}}{D_{ij}^{\gamma}} \]

For outmigration from Katrina affected areas:

\[ m_{i,t} = \alpha + \gamma D_{i} + \beta_1 P_{i,t} + \beta_2 {Katrina} + \beta_3 \mathbf{X_{i,t}} + \varepsilon_{i,t}\]

• \(i\) indicates county destination and \(t\) indicates year
• \(m\) is migration measure
  • Raw flow, inverse hyperbolic sine, share of outflow, and indicator of positive migration
• \(D\) indicates distance from centroid of affected area
• \(P\) is the population for destination county
• \({Katrina}\) is dummy variable for 2005
• \(\mathbf{X_{i,t}}\) are economic characteristics: unemployment rate, median rent, average pay

Model entails destination's distance from New Orleans, its population, a Katrina indicator for 2005, and economic indicators.

• Panel of migration outflow from affected counties from 2000 to 2010
• Start with baseline model of New Orleans area destinations
• Interact explanatory variables with Katrina variable
• Attribute change in effects to the disaster response

• Katrina's impact on outmigration increased the outflows in both volume and breadth, but the distribution of appears unaffected
• Katrina disaster migrants do not migrate as far from New Orleans than in non-disaster years and destination population became less important
• Economic variables indicate mixed results with respect to migration decision of Katrina migrants
• Migration decisions do not appear to mitigate risk as measured via distance
• Data limitations constrain the ability to determine the degree to which migrants return to New Orleans

Questions or comments?

• email: dinterman.1@osu.edu