Playing with Fire? Testing Moral Hazard in Homeowners Insurance Valued Policies
64 Pages Posted: 17 Apr 2017 Last revised: 14 Aug 2019
Date Written: April 15, 2017
Insurance policy design and regulation continually grapples with moral hazard concerns. Yet these concerns rest largely on theory-based assumptions about how rational economic actors will respond to financial incentives. Advances in behavioral economics call these assumptions into question.
This Article conducts an empirical test of moral hazard in homeowners insurance markets. Eighteen states’ “valued policy” laws require more generous compensation by insurers for certain total house losses. I test the moral hazard prediction that fire rates will consequently be higher in these states than in others. Using a private insurance database on the cause of loss for over four million residential insurance claims from 2002 through 2011, I find that, surprisingly, loss rates are significantly lower in valued policy states, not higher. I also use Louisiana’s unexpected elimination of these laws as an additional means to assess the laws’ effects. As before, fire rates are significantly higher when economic incentives appear lower.
These results are inconsistent with standard moral hazard predictions, but I demonstrate how they are consistent with a broader conceptualization of moral hazard theory. First, the results show the importance of recognizing policyholders’ responsiveness to irrelevant factors that they nevertheless believe will affect their insurance payments, like housing prices, rather than the low-salience economic factors that truly determine these payments, like valued policy laws. Second, the results show how focusing exclusively on policyholder behavior misses how other actors, like insurance companies, also adjust to mitigate or even entirely eliminate moral hazard considerations.
Keywords: Insurance, Moral Hazard, Arson, Fire, Valued Policies
JEL Classification: K12
Suggested Citation: Suggested Citation