Demand for Windstorm Insurance Coverage and the Representative Heuristic

With recent research suggesting a behavioral explanation for changes in demand for insurance (e.g., Volkman-Wise (2015)), we study homeowners' demand for windstorm insurance in the wake of losses. In particular, we examine the representative heuristic's impact on the demand for homeowners' insurance, which provides coverage for windstorm losses, in Florida (U.S.). Under this paradigm, individuals underweight prior probabilities and overweight posterior probabilities. This results in an over (under)weighting of the probability of a loss from a disaster in the event (absence) of a disaster. Using data for new homeowners' insurance, purchases in Florida's residual market between 2005 and 2007 (a period of high hurricane activity), we find, subsequent to losses, the demand for coverage limits and the number of policies sold both increase. Further, we find that this effect attenuates as the losses become further away in time. That is, more recent losses have a stronger effect on demand. This attenuation of the demand is also consistent with the representative heuristic.