European bank regulations transform balance sheet management into a constrained combinatorial optimisation problem, yet supervisory stress tests do not, on their own, induce banks to redesign their portfolios. Drawing on a staggered difference-in-differences analysis of 205 European banks across five EBA stress-test cohorts (2016–2025), this paper examines the causal effects of stress-test participation on provisioning, portfolio reallocation and lending dynamics.

The research finds that stress-tested banks reduce provisions by 0.2 to 0.4 percentage points, yet show no systematic shift toward lower-risk asset classes. Safer banks cut lending more aggressively than riskier ones, echoing U.S. findings that supervisory discipline operates through scrutiny and signalling rather than hard capital mechanics. The paper sets out a five-step operating model, combining loan sales, cash securitisations and Significant Risk Transfer with an agentic AI analytics stack, to convert supervisory signals into explicit, repeatable portfolio decisions.

Read the full article here: Bank Balance Sheet Optimisation After Stress Tests.