The Risky Business of Ring-Fencing

Abstract

This working paper examines the impact of imposing multiple ring-fencing regimes on bank safety, including a ‘prisoner’s dilemma’ effect that can arise from host country incentives. It discusses these issues first on a qualitative basis to establish intuition, and then quantifies the risk of different structures via a Merton-style option model. The results suggest that there is a material advantage (i.e. reduced probability of local failure) for a jurisdiction that ring-fences its subsidiary if other jurisdictions do not match that decision. This host jurisdiction benefits from both (i) local capital and (ii) the ability to tap a large central reserve. However, if other jurisdictions adopt similar ring-fencing policies, then the benefit of a pooled ‘central reserve’ capital is voided and all jurisdictions become worse off. Under our model, we find that failure risk increases by a large multiple if ring-fencing becomes pervasive. In some cases, the increase in risk is 5x or more, compared to a structure where internal capital is fully mobile. This is caused by ‘misallocation risk’ – the risk that a bank has enough capital resources overall, but cannot get those resources to the right subsidiary in time to avoid a local failure. The paper concludes with suggestions on possible policy alternatives that could mitigate these issues. We consider a few hybrid cases, which suggest that a ‘partly ring-fenced bank’ can mitigate a substantial amount of misallocation risk if the central reserve is significant and not committed too early. We then explore alternatives which build on the Financial Stability Board’s post-crisis bank resolution architecture. Large amounts of ‘bail-in’ capital are now being issued by the biggest global banks, and these resources can fund an alternative to the hard legal entity bankruptcies (or emergency bail-outs) seen in the 2008 crisis – events that appear to motivate many of the recent ring-fencing initiatives. We conclude with a ‘straw man’ proposal that aims to produce a more resilient overall outcome for the group while addressing legitimate host concerns. We believe the paper is relevant for the ongoing policy debates on bank structure, capital allocation and resolution.