Working Papers

This paper proposes the use of synthetic training data generated by large language models
to improve machine learning SDG classifiers. It shows that supplementing existing training data with
synthetic data produced by the ChatGPT tool improves the performance of the SDGClassy classifier.
This addition of synthetic data is especially useful in building SDG classifiers given the limited availability
of properly labeled data and the complex, interconnected nature of the SDGs. Synthetic data thus enables
more effective machine-learning applications in this context.

This paper investigates the role of trade costs in exporter dynamics in Africa. In comparison to exporters from other regions, African exporting firms are fewer, smaller and relatively less diversified in terms of products and destinations. African countries also display the highest rates of entry, exit and turnover of exporting firms, exporting products and export destinations. This suggests that Africa’s exporting activity is volatile and subject to a lot of experimentation, with exporters having difficulties in maintaining trade relationships. The analysis also confirms that trade costs are a crucial factor in explaining exporter performance in Africa vis-à-vis other regions, but also…