Do global banks need restructuring? 

How should prudential regulators deal with global banks that are 'too big to fail'? Many see bank resolution as the key element in dealing with this challenge (FDIC and Bank of England 2012, Financial Stability Board 2014). The main idea is that global systemically important banks (G-SIBs) should issue 'total loss absorbing capital' (TLAC) in the form of subordinated long-term debt or equity. These securities would be issued to absorb losses and recapitalise the institution in resolution, with minimal disruption to the bank’s operations and without public support.

But what should these resolution frameworks look like? More importantly, would they work? Much hinges on this question, given that there are currently around thirty G-SIBs, with total exposures equal to more than 75% of global GDP in 2014.

Global and local

In our recent work (Bolton and Oehmke 2018), we provide a framework to assess the key trade-offs in cross-border resolution of global banks. The main difficulty in designing effective resolution regimes is the mismatch between their global nature and the national scope of the regulators charged with carrying out their resolution.

Therefore, when designing resolution regimes, we need to consider the political constraints and incentives of national regulators. Drawing on a framework based on standard corporate finance and banking theory principles, we show that conducting a single resolution of the top holding company of the G-SIB – a single-point-of-entry (SPOE) resolution – is efficient in principle. It economises on TLAC and preserves the global bank as a whole.

But this type of global resolution is not always compatible with the interests of national regulatory authorities. When this is the case, resolving a global bank along national lines – a multiple-point-of-entry or (MPOE) resolution – is generally better.

Ex ante and ex post

National interests are important both ex ante (when setting up a resolution regime) and ex post (when the resolution takes place). From an ex ante perspective, a global SPOE resolution usually involves expected transfers from one jurisdiction to another. The regulator in the jurisdiction that is more likely to make transfers may object to setting up a global SPOE resolution framework. Even though a global resolution would be efficient, national regulators that care more about outcomes in their own jurisdiction – for example, because of political economy considerations – would prefer setting up a national resolution regime in order to limit transfers to other jurisdictions. Asymmetries across jurisdictions raise challenges in setting up a global SPOE resolution regime. These are similar to the challenges encountered by EU countries in setting up an EU-wide deposit insurance scheme.

The ex-post difficulties with a global SPOE resolution are perhaps even more serious. National regulators may decide not to honour their commitment when the resolution is carried out. In particular, if the cross-jurisdictional transfer necessary to successfully carry out an SPOE resolution was too large, national regulators would prevent those transfers by ring-fencing assets in their own jurisdiction.

This is the worst of all outcomes, likely to cause a financial panic similar to that following the collapse of Lehman Brothers in 2008. The possibility of an ex-post breakdown needs to be considered when resolution regimes are designed. If regulators know that a global SPOE resolution would not be incentive-compatible ex post, then it would be better to plan for separate resolutions in different jurisdictions using an MPOE resolution.

What determines which type of resolution is appropriate?

Whether a global SPOE resolution can be carried out successfully depends on the risk profile and the operational structure of the global bank in question.

The bank’s risk profile matters because the size of the transfers required during resolution is a function of the correlation of the bank’s cash flows and asset values across jurisdictions.

The bank’s operational structure matters because the incentives for national regulators to make the transfers required in conducting a global resolution depend on how costly it is for them to revert to a resolution along national lines, and break up the bank.

Our analysis shows that credible G-SIB resolution is not one-size-fits-all. Resolution should consider a bank's risk structure and operational complementarities across different jurisdictions. A global SPOE resolution would be more likely to succeed if there were large operational complementarities across jurisdictions that would be lost in case of a break-up through a national resolution. That incentive compatibility could sometimes be restored by following a hybrid approach, in which some of the bank’s TLAC would be pre-assigned to a particular jurisdiction before the resolution took place. Such pre-positioning would reduce the need for ex-post transfers, restoring incentive compatibility, but it would also reduce the benefits of resource-sharing in resolution.

The effect on incentives

We also investigate how the adopted resolution model (SPOE or MPOE) would affect the incentives within the global bank. Equity serves a dual role: outside equity (or long-term debt) serves as TLAC, but inside equity is important for the incentives of bank managers.

Our model shows that a global SPOE resolution would affect incentives for the global bank’s operating subsidiaries in two ways. Relative to MPOE, SPOE would dampen the incentives, because cash flows generated in one jurisdiction would sometimes be transferred to plug a hole in the other jurisdiction. But since SPOE economises on loss-absorbing capital, SPOE resolution might allow the bank to retain a larger inside equity stake, which would allow the bank to provide stronger incentives to affiliate managers. The second, positive effect would dominate if the diversification benefits from a global SPOE resolution were sufficiently large.