Insuring replaceable possessions

Equivalencebetween insuring income and what is bought with income is commonly assumed. It seems to be implicitly held that uninsured but replaceable goods will always be replaced if they fail. This does not follow. People may have difficulty coming up with the money to pay for a replacement out of pocket. Also, the income effect of a loss may mean that replacement is not worthwhile. We show that as a result, equivalence breaks down. Both theory and evidence are provided. Implications include a tendency of empirical papers to overestimate risk aversion, a reason why demand for insurance increases with income, and the mistaken attribution of preference inconsistency.