Comment

Bergin and Lin pick up a venerable distinction that, poorly represented with disaggregated trade data now accessible, has become the dernier cri of trade theory. The general issue of the new literature is what influences the division into intensive and extensive margin of the growth of a country’s share in the global exports going to another country. The specific issue raised in a subset of papers like the one here that is of interest is how, and then why, that division is influenced by the exchange rate regime between pairs of countries. This comment first conveys the essence of the historical distinction and then shows how the data currently used fail to reflect that distinction. This raises the question of what a contemporary implementation of the classical concept, largely preserved in Bergin and Lin’s model but not in its empirical implementation, could imply. It then discusses the absence of substantial uncertainty about future real exchange rates in the model, which detracts from the paper’s main theme. Other aspects that are crucial to the industrial organization of cross‐border trade, such as foreign direct investment (FDI) and trade in components, also are missing from the model. This makes it difficult to test hypotheses with it or to account for its findings. One crude validation test could be to check on the growth of trade shares for Canada and Mexico with the United States since the Canada‐ U.S. Free Trade Agreement 1988 or the NAFTA (North American Free Trade Agreement) 1994 and examine how this growth has been divided between the extensive and intensivemargins. My hunch is that the expansion of the bilateral trade shares of these countries is about as large and as concentrated at the extensive margin as Bergin and Lin estimated for