Amenities or disamenities? Estimating the impacts of extreme heat and wildfire on domestic US migration

Abstract

For decades, Americans have generally been moving away from the Northeast and the Midwest toward the South and West. Mountainous and desert areas in the West and Southwest have been especially popular internal migration destinations. However, these same areas are particularly susceptible to the increasing impacts of climate change on extreme heat and wildfire. This paper investigates relationships between disasterlevel fire events and extreme heat on county-level migration in the USA using annual panel data for the years 1990-2015. Are wildfire and extreme heat associated with changes in migration patterns? And, are relationships between such "disamenities" and migration stronger in counties with closer ties to environmental amenities? Fixed effects and random effects regression analysis show that both experiencing a disasterlevel wildfire and extreme heat in the prior year were associated with reduced migration. Relationships are stronger in counties rich in natural amenities and outdoor recreation, suggesting that in places known for their environmental amenities, shifts toward environmental disamenities may be particularly impactful.

Effects of Wildfire Destruction on Migration, Consumer Credit, and Financial Distress

“…These findings contrast with prior research on non-wildfire rapid-onset environmental hazards, which has generally found that such events have little effect on migration. Our conclusions instead affirm previous research on a smaller subset of FEMA disaster-declared wildfire disasters, which has suggested that such fires are associated with heightened out-migration (Winkler and Rouleau, 2020).…”

“…For fires ranging from 1 to 17 buildings de-stroyed, we instead observe relative immobility -no changes to in-migration or out-migration. This suggests that Winkler and Rouleau's hypothesized "disamenity shifts," wherein environmental characteristics previously considered by residents to be amenities instead come to be considered disamenities, are not occurring to any degree that has affected migration patterns among most residents living with wildfires (Winkler and Rouleau, 2020). This finding should temper any overstatements of mass migration away from fire-prone regions at present.…”

“…Similarly, the impacts of natural disasters on individuals' finances have been explored for each type of disaster except wildfires. This gap in the literature is just beginning to be filled; within the past year, scholars have documented wildfire-related evacuations (Jia et al, 2020) and the migration impacts of different subsets of extreme wildfires (Sharygin, 2021;Winkler and Rouleau, 2020). We build on these efforts in an attempt to develop a more systematic understanding of how wildfires affect migration patterns and household finances, drawing from a full population of all documented wildfires in the United States.…”

“…Many have posited a future of large-scale mobility in response to amplifying climate impacts (Lustgarten, 2020;Rigaud et al, 2018;Hauer, 2017), and there is indeed evidence of preemptive movement away from environmental hazards (Keenan, Hill, and Gumber, 2018). Analysis of disaster-scale wildfires further suggests that the most extreme wildfires may be associated with increased out-migration (Winkler and Rouleau, 2020). Yet, there are also reasons to expect that wildfires will not cause large-scale movement out of fire-affected regions.…”

Population responses to environmental change: looking back, looking forward

“…Additionally, panels of population-environment data can be constructed from administrative data. For example, also in this issue, Winkler and Rouleau (2021) put together a panel consisting of 25 years of observation on almost 3000 US counties from Census data on inand out-migration, NOAA data on extreme heat, Federal Emergency Management Agency (FEMA) data on wild-fire related disasters, US Department of Agriculture (USDA) data on amenities, Bureau of Economic Analysis (BEA) data on wages and salaries, and Bureau of Labor Statistics (BLS) data on unemployment. Clearly, data availability is not the limiting factor it once was.…”