Knowledge mapping of model risk in banking

For years, bank management has relied on mathematical, statistical and financial models, which increasingly expose banks to model risk. The latter is also extended by the phenomenon of innovation: machine learning, artificial intelligence and big data make the models more and more sophisticated and difficult to manage. This study aims to clarify how the literature on model risk is evolving through a bibliometric survey to understand state of the art and identify the discussion topics, open questions and challenges for the future. The study results show that the literature on model risk is still quite young and sparse. The problems to be solved are conceptual, computational, and organizational. The considerations made lead to the question of whether adding further complexity to model risk management is a solution or whether, on the contrary, it creates new model risks.