Our analysis uncovers a large and systematic contraction in the score of EU G-SIBs at year-end. We show that several G-SIBs repeatedly lower their scores to an extent that they reduce their capital requirements. Moreover, a few banks appear to have avoided G-SIB designation altogether in some years. G-SIBs reduce their score in several ways, for instance by temporarily cutting back on intra-financial sector linkages and derivatives positions. G-SIBs with stronger capital ratios and those subject to higher national capital surcharges window dress less than other G-SIBs, which underscores the importance of regulation in banks’ balance sheet decisions. Overall, our findings argue in favour of moving away from using point-in-time data in regulatory requirements and making greater use of averages. They also highlight the importance of supervisory judgment in the assessment of G-SIBs. This could help address banks’ window dressing and mitigate any associated adverse impact on financial markets.
The full paper is available at https://www.bis.org/publ/work960.pdf