Studies measuring barriers to firm growth assume economies are closed, ignoring information on firm exports. We argue that this information is key to interpret data and improve the accuracy of model predictions. To show this, we develop a dynamic model with export and domestic barriers. We show theoretically that the closed economy model under-estimates barriers and amplifies counterfactuals. By calibrating the model to a set of European countries, we find that quantitatively this matters: for example, the closed economy fails to see that Italian firms are very efficient exporters but lousy innovators, and instead conclude that they are mediocre innovators. In terms of predictions, the closed economy model delivers an elasticity of welfare to innovation costs between 31 and 64 percent larger than the open economy model.
