This paper investigates whether firms become more efficient after becoming exporters, and whether exporting generates positive spillovers for domestically-oriented producers. Using firm-level panel data from Colombia, Mexico, and Morocco, the authors analyze the causal links between exporting and productivity. They find that relatively efficient firms are more likely to become exporters, but firms' unit costs are not affected by previous export market participation. This suggests that the efficiency gap between exporters and non-exporters is due to self-selection rather than learning-by-exporting. Additionally, the authors find some evidence that exporters reduce the costs of breaking into foreign markets for domestically oriented producers, but they do not appear to help these producers become more efficient. The authors develop a model of export participation with learning effects, assuming monopolistic competition and constant marginal costs. They find that firms that begin to export exhibit changes in the stochastic process that governs their productivity growth. However, when analyzing actual data, they find that plants that become exporters typically have high productivity before entering foreign markets, and their relative efficiency does not systematically increase after foreign sales are initiated. In some instances, the relatively strong performance of exporters traces to high labor productivity; in other instances, it is due to relatively low costs of intermediate goods. The authors estimate a reduced-form version of the theoretical model to test whether exporting generates external benefits for other firms. They find that export market participation generally depends upon past participation and (weakly) upon past average variable cost (AVC), as implied by the model. However, conditioning on capital stock and past AVC realizations, current AVC does not depend negatively upon previous export market participation, as implied by the learning-by-exporting hypothesis. Thus, the conclusion suggested by their descriptive analysis is borne out by formal Granger causality tests. The authors also find some support for the hypothesis that a firm is more likely to export if it belongs to an export-intensive industry or region, but they find little evidence of any associated productivity gains. Overall, the results suggest that learning-by-exporting is not a significant factor in improving productivity, and that the efficiency gap between exporters and non-exporters is due to self-selection rather than learning-by-exporting.This paper investigates whether firms become more efficient after becoming exporters, and whether exporting generates positive spillovers for domestically-oriented producers. Using firm-level panel data from Colombia, Mexico, and Morocco, the authors analyze the causal links between exporting and productivity. They find that relatively efficient firms are more likely to become exporters, but firms' unit costs are not affected by previous export market participation. This suggests that the efficiency gap between exporters and non-exporters is due to self-selection rather than learning-by-exporting. Additionally, the authors find some evidence that exporters reduce the costs of breaking into foreign markets for domestically oriented producers, but they do not appear to help these producers become more efficient. The authors develop a model of export participation with learning effects, assuming monopolistic competition and constant marginal costs. They find that firms that begin to export exhibit changes in the stochastic process that governs their productivity growth. However, when analyzing actual data, they find that plants that become exporters typically have high productivity before entering foreign markets, and their relative efficiency does not systematically increase after foreign sales are initiated. In some instances, the relatively strong performance of exporters traces to high labor productivity; in other instances, it is due to relatively low costs of intermediate goods. The authors estimate a reduced-form version of the theoretical model to test whether exporting generates external benefits for other firms. They find that export market participation generally depends upon past participation and (weakly) upon past average variable cost (AVC), as implied by the model. However, conditioning on capital stock and past AVC realizations, current AVC does not depend negatively upon previous export market participation, as implied by the learning-by-exporting hypothesis. Thus, the conclusion suggested by their descriptive analysis is borne out by formal Granger causality tests. The authors also find some support for the hypothesis that a firm is more likely to export if it belongs to an export-intensive industry or region, but they find little evidence of any associated productivity gains. Overall, the results suggest that learning-by-exporting is not a significant factor in improving productivity, and that the efficiency gap between exporters and non-exporters is due to self-selection rather than learning-by-exporting.