Banks' window-dressing of the G-SIB framework: causal evidence from a quantitative impact study

Banks' market activity is commonly observed to contract around period-end dates. This behaviour by banks, known as "window-dressing", has micro- and macro-prudential implications, as well as potential repercussions for financial stability and the operationalisation of monetary policy. However, data limitations have constrained the capability of studies to attribute incentives for this behaviour to specific policies, restricting policymakers' scope to impose on banks costly mitigating reporting reforms.

Exploiting a novel and uniquely extensive bank-level dataset, this study employs a difference-in-differences empirical strategy to test whether the response of banks to the G-SIB framework – a key component of the Basel III macroprudential reforms – directly contributes to window-dressing behaviour. It finds causal evidence that it does. These results suggest that banks' attempts to lower their G-SIB scores may be a significant driver of year-end window-dressing behaviour and highlight the potential broader implications of this behaviour on market volatility.