Search

This paper develops a revealed preference method to set-identify the degree of price inattention at the individual level, in a non-experimental, multi-good setting. Necessary and sufficient conditions for consistency with a particular type of price inattention are derived, and used to recover individual-specific sets from household scanner data. This price inattention model significantly outperforms its standard counterpart for most specifications of price inattention considered. The minimum and maximum degrees of inattention are generally small, but vary considerably across households and have significant correlations with household characteristics. These results provide further evidence that heterogeneity in price inattention is empirically important.