The SIPA Capstone team received a project to analyze Financial Market legislations enacted by the European Union (EU) Parliament from its client, LabEx-ReFi. A total of 69 legislations were identified by the client to be analyzed. The outcome of this analysis was to understand the level of discretion given to various agencies, to the member states and to the European Commission from these regulations and the types of changes that have occurred over time. It was decided that the method of analysis would follow that of O’Halloran et al paper entitled, “Data Science and Political Economy: Application to Financial Regulatory Structure” (2017), which used a two-pronged approach.

The team first manually coded, i.e., quantified the amount of delegation and constraints within each of the 69 legislations. This data was then used to calculate a discretion index over the time of observation. The second part of the analysis used Natural Language Processing (NLP) techniques and visualization software to better understand quantitatively and from an unbiased perspective, how these regulations have transformed over time. Using the software “CorText,” the team uploaded the text corpus of the financial regulations and then perform scripted analysis of large quantities of text.