Worldwide, the solution to an efficient mechanism for the resolution of significantly important banks has been bail-in, presenting several advantages over bailouts: it fosters commitment, it protects tax-payers, and it avoids implicit guarantee distortions. However, the implementation of bail-in, which implies loss allocations to banks’ bond-holders, presents significant challenges for global banks since they have branches and subsidiaries in many countries and are subject to regulatory requirements in multiple jurisdictions. The resolution of these problems involves allocating losses across countries, and national regulatory authorities might fail to coordinate since each has an incentive to protect national bond-holders.
Two possible regimes: Single point of entry or multiple points of entry
To address such coordination issues, the Financial Stability Board had suggested enacting bail-in of global banks under two possible regimes: a single point of entry regime, or a multiple point of entry.
In the case of single point of entry, the regulatory authority of the parent holding has statutory power over the entire banking group; under multiple points of entry, both authorities intervene locally. Multiple points of entry therefore is uncoordinated by design, whereas – in principle – single point of entry should solve coordination problems by inducing the resolution authority of the parent holding to internalise cross-country spillovers and externalities. However, in practice the single point of entry may prove unfeasible as the regulatory authority of the local branch might be tempted to intervene despite the original agreement that they would accept the lead of the home country.
Although this new global framework for resolving cross-border banks is still under construction, such coordination issues are already having a major impact on the way national regulators interact across jurisdictions as well as on bank’s business models and internationalisation strategies.
Optimising global banks and national regulators under various regimes
The analysis of coordination in prudential regimes is not new. For instance Calzolari et al. (2015) study the coordination in prudential policy with special focus on the efficiency of the design of the European Single Supervisory Mechnanism. Dell’Ariccia and Marquez (2006) assess the role of regulatory competition on financial integration. No paper has so far analysed coordination externalities for resolution policy. In a new paper (Faia and Weder di Mauro 2016a), we propose a theoretical framework to analyse these regulatory innovations and their intended, as well as unintended, consequences.
We do so using a game-theoretic approach and assessing the welfare consequences of different regimes. We consider official regimes – single and multiple points of entry – and the failure of a single point of entry regime (what we call ‘single point of entry non-cooperative’). Importantly, banks and regulators in our model are both optimisers, in that banks solve portfolio optimisations subject to equity requirements and subject to the regulations around maximising social welfare trading off the interests of the bond-holders against the interest of the equity holders. Indeed, increasing the fraction of ‘bail-in-able’ bonds raises ex post losses to bond-holders, but facilitates the satisfaction of the equity constraint, thereby reducing equity holder involvement into resolution. The presence of a social welfare function allows us to draw conclusions about the efficiency of the regimes and about the welfare costs of mis-coordination.
Notice that cross-country spillovers in our model arise because of a shared liability effect and because of an asset internationalisation effect. Global banks raise liabilities in several markets and use them to satisfy group-wise capital requirements. When a national regulator attempts to protects national bond-holders by reducing the loss they bear, this indirectly affects the equity requirement of the global group, and hence global financial stability. Finally, banks invest in assets across the board. Hence equity requirements on foreign asset holding feed back on the optimal amount of bail-in losses of domestic bond-holders.
Consequences for welfare, financial integration and bank business models
We assess the results along two dimensions: efficiency and effects on financial integration. Indeed, once the optimal regime has been chosen by the regulatory authority, the banks in each country react to it by deciding whether to remain global or not and also by endogenously adapting their international business model (branch versus subsidiary).
We find that single point of entry regimes are the most efficient in terms of minimising overall welfare losses (total of bondholders and equity holders in both countries), as intended. This is in line with the explicit preference regulators and the majority of global banks have expressed for this regime (see the section on living wills below). However, we also show that single point of entry resolutions can have unintended consequences as they may foster financial disintegration in previously segmented markets.
The multiple points of entry approach generates higher bond holder losses, in particular if the bank is funded primarily in the home country, i.e. it has a high degree of home bias in its liabilities. However, multiple points of entry can be more efficient in the case of decentralised banking groups, characterised by a preponderance of local funding in the foreign countries. Again, this finding is in line with the expectations of regulators as well as with the intentions expressed in living wills of banks with this type of international retail model.
Given that single point of entry is emerging as the preferred choice of most home regulators and global banks, we focus on a third equilibrium by relaxing the central condition for the efficiency of single point of entry, namely that the foreign authority accepts the lead of the home country (i.e. they both cooperate). We show how non-cooperation affects the equilibrium loss allocation and internationalisation if cooperation breaks down – bond holders’ losses increase and banks react by reducing cross-border exposures and by increasing subsidiarisation. This might explain part of the observed process of global banks' subsidiarisation. If the fear that cooperation under a single point of entry will not be sustained in crisis is indeed leading to inefficiently large risks and to financial disintegration, this would clearly be an unintended consequence of the new bank resolution architecture.
A redeeming factor for global regulation emerges from a final feature of our model. Higher capital requirements at home and abroad (in line with the new requirement for much higher total loss absorbing capital in systemic banks) work as discipline devices, thereby reducing the ex post overall default losses. This also implies that spillovers are milder, hence the distinctions between the different regimes are blurred and welfare losses reduced in all regimes.
Living wills of global banks are in line with the findings
Our model speaks about normative aspects of the financial architecture regimes. But we also examine the emerging landscape of cross-border resolution, which is now possible because regulators worldwide require that global banks submit a plan that explains how the recovery or resolution would take place in the event of severe financial distress or failure of the bank. Such recovery and resolutions plans, often called ‘living wills’, have to be submitted for approval to various regulators worldwide and the cornerstone of all plans is the designation of the global banking group as operating under the single point or multiple points of entry regime.
We examine the living wills of global banks with more than $100 billion in assets that submitted to the Fed in 2015 to find the choice of a global resolution regime.1 Table 1 shows that the majority of global banks are to operate under the single point of entry regime; only HSBC and BBVA are designated as multiple entry point groups.
Table 1. Living wills (recovery and resolutions plans) of global banking groups
Notes: Summaries published by NY Fed (www.federalreserve.gov/bankinforeg/resolution-plans.htm). Included: banks holding companies with more than $100bn in assets that submitted recovery and resolution plans to the NY Fed in 2015. Not included: financial groups (AIG, Prudential Financial, General Electric Capital Corp.); Wells Fargo and Bank of NY Mellon (bridge bank). Total assets as reported in RRPs/Annual Report (AR). Comparability is restricted due to different accounting standards.
The table also shows that the choice of single or multiple points of entry is not correlated with the size of the bank. HSBC and BBVA are, respectively, the largest and the smallest among those banks in terms of total assets. Instead, the choice of multiple points of entry appears to be related to the highly decentralised, retail-based business models of the two banks. For instance, BBVA has expanded internationally (mainly in Latin America) through the acquisition of local retail banks, which continue to operate with local funding and quite independently from headquarters. For such a bank, a multiple entry point strategy seems a better fit since the scope of strategic interaction is limited. This observation is well in line with results of our model, which show that only for a low degree of home bias in liabilities (ex ante highly decentralised business models) will multiple points of entry regimes dominate. For more centralised global banks, a single point of entry is more efficient – provided that authorities are able to coordinate.
Calzolari, G, J E Colliardy and G Loranth (2015), “Multinational Banks and Supranational Supervision”, Bundesbank website.
Dell’Ariccia, G and R Marquez (2006), “Competition among regulators and credit market integration”, Journal of Financial Economics 79(2): 401-430.
Faia, E and B Weder di Mauro (2016a), "Cross-Border Resolution of Global Banks: Bail in under Single Point of Entry versus Multiple Points of Entry’’, CEPR Discussion Paper No. 11171.