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.

The pandemic-induced global economic crisis has contributed to the re-emergence of sovereign default risk, especially for emerging and developing economies, and has directed attention to the impact of the institutions that are tasked with attempting to predict defaults: the international credit rating agencies. This paper describes four main challenges posed by credit rating agencies, especially from a developing and emerging economies perspective: potential bias in ratings, pro-cyclicality of ratings, governance issues and conflicts of interest, and incorporation of climate risk. It concludes with potential policy solutions addressed at ratings agencies, regulators, and policy makers.…