Highlights :

 Abstract :
Firm level data exhibits that new exporters tend to start small, a large fraction of these drops out by the second year of exporting, and the survivors expand rapidly. To take into account this stylized fact, we propose a theory of firm behavior that assumes demand uncertainty about the profitability of exporting in a new market. The firm can postpone paying the sunk cost of full-scale entry and test the market by observing individual sales to a few consumers. The firm optimally chooses the experimentation intensity, as well as the exit/entry policy. Applying Bayesian econometric techniques, we structurally estimate the model using French firm-level export data. A given geographical regions is viewed as a target market,and countries within the region as consumers. The estimate of the sunk cost is higher than in a model where the sunk cost cannot be postponed, like Melitz (2003). We also perform counterfactual simulations (exchange rate, sunk cost and experimentation cost).

 Keywords : Demand Uncertainty | Optimal Experimentation | Heterogeneous Producers | New Exporter Dynamics | Structural Estimations | Bayesian Methods

 JEL : D21, D83, F14, C11, C33