The economic consequences of flood alerts: evidence from employment outcomes

Abstract

Floods are some of the costliest and most common natural hazards in the United States. As the frequency and intensity of extreme weather events increase, flood alerts have become an important tool for reducing damages and saving lives. Previous studies show that early warning systems can reduce economic losses and protect lives, but their effectiveness depends on alert accuracy and public response. Most of the existing research has focused on how false alarms and missed alerts affect trust and compliance, especially in the context of tornadoes. Much less is known about their economic impacts, particularly for floods, despite the fact that issuing an alert involves tradeoffs. False alarms may generate unnecessary costs, while missed events can lead to unanticipated disruptions. Understanding the labor market impacts of these inaccuracies is important for agencies like the National Weather Service, which must balance risk communication under constrained resources. This thesis investigates how flood alerts, false flood alarms, and missed flood events disrupt county-level employment outcomes in the United States. To answer this, I construct a county-by-month panel dataset from 2007 to 2023, combining data from the NOAA Storm Events Database, the Iowa Environmental Mesonet, and the Bureau of Labor Statistics. Using a two-way fixed effects model, I estimate how different combinations of alerts and flood events influence employment levels in total, as well as construction, and leisure & hospitality sectors. The results show that flash flood days accompanied by both a warning and a notification are associated with small but statistically significant declines in employment. These effects are most pronounced in the leisure and hospitality sector, followed by construction. I find no evidence that false alarms or missed events have a significant impact on employment. Additional analyses reveal that effects vary by rural and urban classification, emerge over several months in some cases, and are robust to alternative alert groupings and proximity-based exposure definitions. These findings highlight the economic importance of alert accuracy and suggest that the type and layering of alerts, as well as local context, influence short-term labor market outcomes.