The language of rules: textual complexity in banking reforms

The implementation of Basel III banking reforms that followed the financial crisis of 2007–08 led to an increase in UK banking regulation from almost 400,000 to over 720,000 words. These reforms have also led to concerns about the complexity of financial regulation. However, the debate lacks clarity on: (1) how to measure this complexity; and (2) the extent to which technology can be used to address it. We restrict our analysis to cognitive costs related to language processing, and construct a new textual dataset of the prudential rules that applied to UK banks before and after the Basel reforms. We use natural language processing and network analysis to calculate complexity measures on this novel dataset. We find that, while the language of individual rules remained stable, suggesting a continuity in drafting style, rules became more interconnected, via longer chains of cross-references. In following these chains, the number of words that a reader had to process starting from a single rule increased from about 600 words to over 25,000 on average (an increase of over 4,000%). We also contribute to developing textual measures that can help identify rules that are complex for humans but require limited interpretation and are better suited for translation into machine-readable code.

Download paper here

We published a blog post outlining the results.

This is a dormant working paper.

BibTeX citation:

@article{amadxarif2019language,
  title={The language of rules: textual complexity in banking reforms},
  author={Amadxarif, Zahid and Brookes, James and Garbarino, Nicola and Patel, Rajan and Walczak, Eryk},
  year={2019},
  journal={Bank of England Working Paper},
  volume={834}
}