摘要：This paper extends recent work by Martin (2005) on estimation of total factor productivity by allowing heterogeneity of markets in terms of competitive pressure. We show that conventional methods of total factor productivity estimation yield biased measures of TFP when firms charge market-specific prices. Specifically, if, as widely believed, exporting price-cost markups are, on average, lower than domestic markups, then traditional total factor productivity measures will substantially underestimate exporter productivity and overestimate the productivity of non-exporters. Crucially, our results provide some possible clues on why empirical studies failed to find conclusive evidence on the learning-by-exporting hypothesis.
