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 argues that SDRs should become a more relevant instrument of international monetary cooperation. This requires transforming them into a pure reserve asset and the IMF into a fully SDR-funded institution. SDRs would then be issued counter-cyclically and treated as deposits of countries in the IMF, which can in turn lend to countries. This approach would correct basic deficiencies of the current global monetary system. Complementary reforms include a substitution account for an orderly and smooth transition from major reserve currencies to SDRs, and the issuance of SDR-denominated bonds as an alternative to other major short-term assets.