"This paper looks at the predictions of a standard heterogeneous firm model regarding the exports of firms across markets in response to a particular trade policy "experiment" and compares these predictions to the data. A unique feature of our data is that it has information on the exports of the same firm to different markets which allows us to look for a new set of predictions of such models. We argue that while certain predictions seem consistent with the data, others are not. We then describe the patterns found in the data and argue that firm and market specific demand shocks help explain a number of these anomalies. These parsimoniously capture factors, like business contacts or networks, or even fashion shocks, that make buyers more attracted to one firm rather than another in a particular market"--National Bureau of Economic Research web site.