Natural disasters often lead to changes in risk management decisions. They can provide new information on a risk, make the risk salient, or draw attention to particular policies or activities. These effects could lead people to invest in risk reduction after a disaster, when they did not consider the threat in fair weather. Research has shown that after flood events, households are more likely to invest in hazard mitigation (see: here and here, for example) and housing prices often drop (see here, here, here, and here). Several studies have also documented an increase in flood insurance purchases after a flood (here, here, here).
When it comes to increases in flood insurance purchases, however, federal disaster aid policy could be playing a role. Recipients of federal disaster aid in a high flood risk area are required to purchase flood insurance. In a recent paper published in Risk Analysis, I estimate the extent to which this requirement explains the observed bump in take-up rates of flood insurance after hurricanes.
Overall I find that when the center of at least one hurricane passes through a county in the previous year the overall take-up rate of flood insurance increases by a little more than 7 percent. This is temporary, however, and the effect dies out by three years after the storm. Presidential disaster declarations increase flood insurance purchases by a similar amount: 6.7 percent. Most of this effect is explained by the authorization of Individual Assistance, the FEMA program that provides post-disaster grants for repair and rebuilding to households, and which carries the requirement that recipients insure. When this program is available, there is an increase in take-up rates of 5 percent.
When I look specifically at which flood policies were required because the household received disaster assistance, hurricanes only lead to a 1.5 percent increase in take-up rates of voluntary (that is, not required by disaster aid) purchases. The requirement that disaster aid recipients insure, then, is responsible for much of the increases in insurance purchases post-hurricane. My co-authors and I found a similar result in Florida: there was only a slight increase in take-up rates in zip codes that received Individual Assistance, attributable to the requirement that disaster aid recipients insure.
This work suggests that policy requirements may be necessary to increase take-up rates of disaster insurance since that accounts for the majority of post-storm jumps in insurance purchase. Indeed, many types of insurance are often not bought without some type of requirement. Banks require homeowners insurance, states require automobile insurance, and the federal government is now requiring health insurance. Where no such regulations exist, very few people insure, as seen with earthquake insurance in California and areas where flood insurance is voluntary. Policies that capitalize on increased interest in a risk after occurrence of a disaster may be a way to require insurance without as much political opposition as requiring people to insure before a disaster. Guaranteeing households keep these policies over time, however, might be more difficult and worthy of future work.