Exploiting disaggregated data on French exporters, we show that firms expand their product scope and geographical presence sequentially. This process of internationalization is uneven over time, exhibiting more volatility early than later in the life cycle of exporters. Specifically, young exporters are particularly likely to exit, and if they keep exporting, to expand at the intensive and sub-extensive margins, doing so by widening product scope within a destination before entering new destinations. We also find that firms' core products are particularly resilient despite being used to "test the waters" when entering additional countries. Existing models of firm export dynamics are not designed to explain these empirical regularities. We argue that they can be rationalized by a mechanism where new exporters are uncertain about the profitability of their products in different markets, but learn from their initial export experiences and then adjust their sales, number of products and destination countries accordingly.

This paper explores the determinants of firm survival in export markets. We build an exporter dynamics model where firms need to pay market-specific sunk and fixed costs to operate abroad and where firm export profitability in each foreign market follows a geometric Brownian motion. Firms also differ ex ante by a constant market-specific profitability shifter. We derive the probability of export survival upon entry in a market and show that it increases with the ratio of sunk to fixed costs and is insensitive to the profitability shifters. Also, we show that the survival probability is unaffected by fixed costs if sunk costs are zero. We take the model to the data using firm-level Argentine export information. We find that survival rates decrease with distance, which the model rationalizes with sunk costs that increase with distance proportionally less than fixed costs. Estimated sunk costs are small. In fact, a counterfactual exercise shows that removing those costs increases aggregate exports by less than 1.5\%. Finally, we also find that survival increases with a firm’s export experience. Analogously to distance, the model’s implication of this empirical result is that experience reduces sunk costs proportionally less than fixed costs.