Research on Emerging Risk Identification and Mitigation Mechanisms in the Digital Transformation of Financial Institutions

Abstract

Digitalisation of banking has introduced new risks, such as cybersecurity, data protection, algorithmic accountability and third-party dependence, which are not covered by the traditional risk indicators. This paper builds a text-based system to identify these exposures and studies their mitigation by analyzing the annual reports of the 20 listed Chinese commercial banks from 2017 to 2023. Dictionary-based indices for the four risk categories are constructed from disclosure narratives, their validity is assessed by distinctness from conventional credit risk and alignment with regulatory milestones, and they are related to a lagged transformation index in two-way fixed-effects panel models supported by wild cluster bootstrap inference. Disclosed exposure more than doubled in the window, driven primarily by data-related risks that surged around new privacy laws, and the composition of these risks varied among institutions. Transformation intensity, however, carries no within-bank association with emerging risk exposure, a null robust to alternative index constructions and an alternative risk measure, and governance moderation through board independence finds no support under admittedly thin identifying variation. The evidence points out that the new risks are industry-wide phenomena, supports a multi-level mitigation system emphasising supervisory coordination and segment-specific technical standards, and leaves governance reform as a question for longer panels.