Return on investment on artificial intelligence: The case of bank capital requirement

•We contribute to the literature by providing an empirical exercise computing banks capital requirement with the help of AI techniques for predicting corporate defaults.•We measure to which extent banks have incentives to invest in AI techniques at the light of capital requirement economy induced by these techniques.•We single out neural networks among the AI techniques to be more likely to meet the regulatory expectation as well as to lead to the strongest RWA reduction. We note that the traditional approach is closed in term of performance.•We document the fact that the level of capital requirement depends on the statistical methodology used by the bank for setting up its internal models. Taking advantage of granular data we measure the change in bank capital requirement resulting from the implementation of AI techniques to predict corporate defaults. For each of the largest banks operating in France we build by an algorithm pseudo-internal models of credit risk management for a range of methodologies extensively used in AI (random forest, gradient boosting, ridge regression, neural network). We compare these models to the traditional model usually in place that basically relies on a combination of logistic regression and expert judgement. The comparison is made along two sets of criterias capturing: the ability to pass compliance tests used by the regulators during on-site missions of model validation (i), and the induced changes in capital requirement (ii). The different models show noticeable differences in their ability to pass the regulatory tests and to lead to a reduction in capital requirement. While displaying a similar ability than the traditional model to pass compliance tests, neural networks provide the strongest incentive for banks to apply AI models for their internal model of credit risk of corporate businesses as they lead in some cases to sizeable reduction in capital requirement.