Firms can increase unit prices by decreasing package content, a practice known as product downsizing. Because consumers tend to underuse size information, they may fail to notice size changes. Downsizing in the pepper industry provides an opportunity to test consumer inattention. We derive a novel method to study inattention to changes in package content and apply it to grocery scanner data. We find that almost all consumers fail to notice change in package content. With full information, consumers would switch to larger packages, but the overall improvement in welfare is small as consumers care more about price than size.
Estimating multinomial models with many fixed effects faces prohibitive computational challenges. We develop a new method based on the Minorization-Maximization (MM) algorithm to address this issue. The method linearizes the logit model, allowing for the use of linear methods for handling fixed effects. We demonstrate the usefulness of our method through simulations and an empirical application to hospital choice. We show that our method also works well if the researcher wants to specify a model with individual-specific coefficients on an explanatory variable, even with large numbers of individuals.