Sample zeros---products available that do not register positive revenues due to the finite sample period---introduce several biases in the computation of the import price index developed by Feenstra (1994). We model sample zeros by considering the discrete choice model of Anderson, De Palma and...

reallocation are halved.

disaggregated trade data from patterns arising mechanically through chance.

Exporters are few-less than one-fifth among U.S. manufacturing firms-and are larger than non-exporting firms-about 4-5 times more total sales per firm. These facts are often cited as support for models with economies of scale and firm heterogeneity as in Melitz (2003). The authors find that the...