Reform of the CMDI Framework – Driving Off With the Brakes On

3.1. The predicament under the CMDI framework: impairing deposits in bank crisis management

Banks are resolvable when they can absorb losses and, if needed, recapitalize themselves while preserving at least their critical functions.^[16] Three components contribute crucially to this end by providing the necessary loss-absorbing and recapitalizing capacity: (i) the MREL instruments and the additional (non-MREL eligible) bail-inable liabilities which are written-off or converted in case of failure, (ii) resolution financing arrangements which provide capital (and liquidity) in resolution when bail-in is not sufficient, and (iii) deposit insurance which funds resolution when covered deposits would have been bailed-in in resolution^[17] or reimburses covered depositors when they are haircut in liquidation. If one of these components is unavailable or underperforms, the pressure for adequate financing falls upon the others. Ultimately, the success of the crisis management efforts can be at risk.

Commentators have voiced concerns that resolution financing in the EU is currently inadequate.^[18] We share this view,^[19] mainly because the current CMDI framework has critical shortcomings.

539Member states’ resolution financing arrangements and the SRF, which mutualized national resolution funds in the euro area, are primarily designed as a recapitalization tool for European banks only once the 8% bail-in of the failing bank’s TLOF has been applied. Moreover, resolution funds can contribute only up to 5% of the failing bank’s TLOF.^[20]

DGSs can also provide resolution financing.^[21] These schemes can contribute to resolution up to the amount they would have to reimburse if covered depositors were subject to internal PSI at the failed institution. In particular, if resolution schemes foresee the application of the bail-in tool, the DGSs are liable for the amount covered deposits would have been written down to absorb losses (not: to recapitalize the bank) were they not exempt from internal PSI.^[22] When resolution transfer tools are applied, the DGS is liable for the losses covered depositors would have suffered had they been treated like creditors ranking equally under ordinary insolvency proceedings.^[23] DGSs’ contributions reflect the protection covered depositors enjoy in case of liquidation and thus avoid violating the ‘no creditor worse-off’ principle (“NCWO”),^[24] also for pari passu creditors.

Moreover, the DGS’s liability should not be greater than the amount equal to 50% of its target level unless member states increase the maximum contribution.^[25] This contribution is further limited by potential recoveries on reimbursed deposits.^[26] First and foremost, however, the super-priority provided to insured deposits makes using DGS financing in resolution very unlikely be540cause the essential hypothetical loss can only occur after all other (lower ranking) liabilities, including uninsured deposits, would have been bailed-in.^[27]

The current CMDI framework may create a scarcity of resolution financing. Figures 1a–1c and 2a–2c illustrate the resolution funding gap that may occur under the waterfall for internal PSI under the current CMDI framework. They distinguish between those member states that rank ineligible deposits, i. e., non-SME corporate deposits, pari passu with other unsecured liabilities on the one hand (Figures 1a–1c) and those member states that rank these deposits higher than other unsecured liabilities on the other (Figures 2a–2c). They show that the balance sheet composition of the failing institution is crucial for effective internal PSI that avoids impairing runnable deposits. Therefore, resolution strategies relying on internal PSI are only credible if the bank has sufficient liabilities (other than uncovered deposits) that can be bailed-in to meet the 8% TLOF minimum requirement without destroying value beyond their nominal value,^[28] and, most importantly, without triggering a run.

541

Figure 1a-1c

– Funding gap with three tier depositor preference (NSMS)

542

Figure 2a-2c

– Funding gap with three tier depositor preference (SMS)

543The looming funding gap effectively compels resolution authorities to ensure the resolvability of banks through the MREL calibration if they intend to resolve banking crises under the CMDI framework without touching runnable deposits.^[29] At the outset, EU law allows for such an MREL calibration. The CMDI framework requires all banks, regardless of size and business model, to meet demanding MREL prescriptions^[30] and, therefore, compels banks to build up buffers of subordinated liabilities^[31] that exceed their prudential own funds requirements if the resolution plan foresees resolving the bank under the CMDI framework (instead of liquidating it in bankruptcy). However, many retail banks, regardless of the soundness of their business models and financing strategies, face severe obstacles in accessing public capital markets at sustainable costs.^[32]

According to empirical evidence provided by the European Commission^[33], in a sample of 187 banks that would go into resolution if they failed as of Q4 2019, deposits in 44 banks in 18 member states would have to bear losses for an aggregate amount of EUR 14.16 bn under the baseline scenario (status quo) to 544reach 8% TLOF and be able to access the national resolution financing arrangements/the SRF. Retail and SME deposits are the predominant source of funding in smaller and medium-sized banks across the EU. Exactly these institutions find it challenging to place MREL instruments on capital markets, with their alternative funding strategies as prototypical deposit-taking credit institutions not being objectionable per se.

3.2. Resolution authorities’ incentives to manage banking crises outside the CMDI framework

The current CMDI framework treats uncovered deposits as bail-inable liabilities, though , in principle, not as MREL-eligible ones. At the same time, access to external resolution funding becomes an option only once a predetermined amount of internal PSI, fixed at 8% TLOF, has been achieved. Depending on banks’ funding structure, this bail-in minimum can require touching on uncovered deposits to tap resolution funding arrangements. The CMDI framework seeks to avoid the need for such a bail-in of uncovered deposits by prescribing an MREL buffer of sufficient quality and quantity that ranks lower than uncovered deposits in the creditor hierarchy.

However, the structure of the EU banking sector, characterized by many small and medium-sized institutions with, at best, limited access to capital markets, makes relying on MREL calibration alone a dubious policy choice (see also below 4.2.2). If resolution authorities impose demanding MREL targets on these institutions, they might trigger an undesirable concentration in banking.

Transfer strategies in resolution could, in principle, reduce the reliance on PSI by activating private sector backstops to stabilise wobbly banks. However, such strategies remain underdeveloped and underutilized in the EU, not least because they also suffer from insufficient resolution funding. Buyers are generally more willing to acquire a failing bank and assume its risks if they only need to recapitalize the institution instead of also covering its losses. Transfer strategies are, therefore, also a more viable option in bank crisis management if the failed acquisition target has sufficient and credible loss absorbing capacity.

The EU legislation presumes that such loss absorbing capacity could come from the bail-in of uncovered deposits. Yet, this presumption may be incorrect. The literature and experience with bank crisis management (see above) suggest that the (rational) expectation to incur losses in bank resolution may trigger bankruns^[34] and entail undesirable welfare losses.^[35] Resolution authorities and political 545decision makers understand that predictable funding gaps (see above 3.1) may prevent achieving resolution objectives better under the CMDI framework than in liquidation. Financial stability would be at risk if uncovered deposits had to be bailed-in in resolution. No such risks exist in liquidation if applicable insolvency laws and state-aid rules allow for a more generous injection of stabilizing government funds (bail-outs) and thus permit decision makers to leave covered and uncovered deposits intact.^[36] Against this background, the existing CMDI framework with its looming funding gap induces resolution authorities to opt for liquidation instead of resolution. Resolution authorities face a dilemma: while the CMDI framework expects authorities to achieve the resolution objectives with the resolution toolkit, the lack of sufficient resolution funding is apt to thwart such efforts. Consequentially, all else equal, policy considerations under the current CMDI framework^[37] would militate in favor of liquidation supported by extraordinary government funding, leading to sub-optimal crisis management choices, financed from public budgets and following widely varying national regimes.^[38] Resolution authorities can accommodate their preferences by already locking-in liquidation strategies in institutions’ resolution plans.

4.1. Background: member states’ reluctance to fully centralize resolution and deposit insurance

The incentives to manage banking crises outside of the CMDI framework (3.2) are likely to persist beyond a transition phase insofar as MREL does not pro546vide a fully viable solution, particularly for capital market remote small and medium-sized banks. Sound banking policy may thus favor legislative amendments that allow tapping other sources of resolution funding, notably if they ensure external PSI. Therefore, the objective of the CMDI reform was to make resolution of retail banks possible without exacerbating run risks. Yet, the reform effort was an attempt to optimize under constraints. Since its inception, the Banking Union was conceived as a three-pillar structure, with the third pillar being a common mechanism for deposit insurance.^[39] However, even though the Commission adopted a legislative proposal already in 2015,^[40] a fully mutualized European Deposit Insurance Scheme (“EDIS”) remained a no-go for several member states over the years.^[41] Retaining national idiosyncrasies like institutional protection schemes allowed them to shield politically well-networked parts of domestic banking systems from the arguably more demanding requirements of the CMDI framework.^[42] Against this background, the Commission initiated the review of the CMDI framework in 2019 with limited scope, signaled in its report on the application and review of the BRRD and the SRMR.^[43]

The political constraints that militate against a full-fledged reform implementing a first-best approach to completing the Banking Union^[44] will likely persist or become even more pronounced in the near future as the opposition to 547further centralization grows in many member states.^[45] Therefore, a critical assessment of the 2023 reform proposal remains worthwhile because it can improve real-world policy choices once the co-legislators carry the CMDI reform project forward during the 2024-2029 legislative term.

4.2. What would the CMDI Proposal have achieved, and what would the remaining shortcomings be?

4.2.1. Enhanced industry resolution funding

The CMDI Proposal aims to increase available resolution funding by introducing two changes to the existing CMDI rules – (i) it amends the creditor hierarchy by canceling the super-priority of covered deposits and treating covered deposits pari passu to uncovered deposit (Article 108(1) BRRD3); (ii) it allows the use of DGS funding as a replacement to eligible deposits that would otherwise have to be bailed-in in a transfer-based resolution strategy (Article 109 BRRD3). The ECON Report takes a similar position in respect of Article 108 BRRD3 but proposes a two-tier depositor hierarchy separating retail and SME deposits from large entreprise and central and regional government deposits. The Council Report, however, proposes that the super-priority of covered deposits is retained in its proposed version of Article 108 BRRD3.

4.2.1.1. Depositor preference

Article 108(1) BRRD3 abolishes the distinction between covered and uncovered deposits in the creditor hierarchy. Instead, it treats all deposits (covered and uncovered as well as eligible deposits of corporates and excluded deposits) as one creditor class that sits above all ordinary unsecured claims and ranks pari passu to the DGS. This amendment, all things being equal, would already permit the use of the DGSs’ funds earlier because loss absorption through internal PSI would not have to burn up uncovered deposits to trigger DGS contributions (see above 3). Therefore, DGS funding becomes more accessible, and the looming funding gap occurs less frequently. Figures 3a-3c illustrate this effect.

548

Figure 3a-3c

– Funding gap and DGS contribution with single depositor preference under the CMDI Proposal

549This amendment is also consistent with the literature on deposit insurance, which argues that the co-existence of deposit insurance and covered deposits’ super-priority minimizes the useability of deposit guarantee schemes.^[46] The policy decision of the EU legislator to protect covered deposits in bank crisis management remains intact since covered deposits still benefit from an exclusion from bail-in and would, in any case, be reimbursed by the DGS within a short period (Recital 39 and 43 BRRD3). From a financial stability perspective, DGSs’ capacity to step in in a resolution scenario and cover deposits without incurring tail effects for the financial system is not straightforward, however – especially considering the funding model of DGSs, which relies heavily on ex post contributions.

The ECON Report confirms the abolition of the distinction between covered and uncovered deposits. However, it proposes a two-tier depositor hierarchy where retail and SME deposits rank pari passu to covered deposits and higher than deposits of large enterprises and central and regional governments. This proposal enhances the protection of retail and SME deposits. It also conforms with existing MREL rules that consider some deposits to be MREL eligible. Yet, compared to the CMDI Proposal, the amendment would also increase the run risks because it requires the bail-in of large enterprises’ and central and regional governments’ deposits before DGS funding can be activated. As a result, the use of DGS funding in resolution would remain unrealistic at banks with significant wholesale and sovereign deposits on their balance sheets (see section 4.2.1.2 below).

The Council Report instead preserves the current super priority of covered deposits and, therefore, largely the current status quo around DGS funding in resolution. It further complicates matters by introducing a four-tier framework: Covered deposits would hold the highest priority, followed by uncovered deposits of SMEs and natural persons, other uncovered (bail-in eligible) deposits with a maturity of less than one year, and other (also bail-in eligible) uncovered deposits with a maturity of one year or more.

4.2.1.2. Use of DGS funding in resolution

Consistent with creating a simpler depositor hierarchy (either single or two-tier depending on the outcome of the legislative negotiations) (see above 4.2.1.1), DGS contributions can substitute for a more comprehensive set of losses predicated on resolution actions. BRRD3 art. 109 specifies the use of 550additional sources of funding that become available due to dropping covered deposits’ super-priority (supra 4.2.1.1).

The leeway for DGS funding in resolution is not materially increased in an open-bank resolution strategy with or without the use of the asset separation tool, i. e. a resolution strategy that aims to resolve the bank as a going concern without exiting the market or transferring its assets to a buyer and relys primarily on the bank’s own resources for loss-absorption and recapitalization. The DGS contribution remains limited to the hypothetical losses covered deposits would have suffered in a bail-in, although DGSs can also contribute to the failed banks’ recapitalization.^[47] Hence, the reforms cannot bridge resolution funding gaps at banks that rely heavily on uninsured deposits and, therefore, have no bearing on run risks similar to those that materialized in spring 2023 at the faltering US regional banks. (Recital 45 BRRD3).

DGS contributions are available beyond covered deposits’ hypothetical losses if resolution authorities pursue transfer strategies (sale of business and bridge institution). The European Commission’s proposed Article 109(1) subpara 1 (b) BRRD3 specifies that DGS contributions to transfer strategies shall consist of (i) the amount necessary to cover the difference between the value of the covered deposits and of the liabilities with the same or higher priority ranking than deposits and the value of the assets of the institution under resolution which are to be transferred to a recipient,^[48] and (ii) where relevant, an amount necessary to ensure the capital neutrality of the recipient following the transfer. Although the exact scope for the use of DGS contributions in transfer strategies remains questionable,^[49] the essential feature of the proposed rule is straightforward: DGS contributions can cover losses and satisfy recapitalization needs that would, absent DGS interventions, create a burden for uncov551ered depositors (and other pari passu creditors).^[50] At the outset, the reform could thus effectively quell detrimental run incentives.

Moreover, DGS funding in resolution can bridge a gap between the failed entity’s MREL capital resources and the 8% TLOF requirement to access national resolution financing arrangements or the SRF (see above sections 2 and 3).^[51] The proposed amendment would open two tiers of resolution funding with external PSI once a failing bank’s internal loss-absorbing capacity beyond deposits is exhausted. It would, in principle, close the funding gap in resolution many retail banks face and mitigate the retail challenge that many banks, resolution authorities and political decision makers currently struggle with. Under the ECON Report, the DGS funding options are broadened because a DGS can also contribute when funding from resolution financing arrangements has reached the 5% cap (see above 3) and no other liabilities than deposits could be bailed-in.

However, the ECON Report’s proposal for a two-tier depositor preference model (above 4.2.1.1) may hinder the availability of DGS funding in resolution. The demise of the US regional banks in early 2023 illustrated that the bail-in of large entreprise and government deposits might generate financial stability risks (depending on the type of bank and jurisdiction in consideration). Therefore, the ECON Report’s proposal places resolution authorities between a rock and a hard place. They can either avoid stability risks by refraining from a bail-in of the respective liabilities, thereby constraining their access to DGS for resolution funding. Alternatively, they could accept the impending perils and impose losses on wholesale and sovereign deposits, thereby gaining access to DGS resolution funding. Although the latter may not always be the dominant strategy for a considerate resolution authority,^[52] the increased possibility of suffering losses on their deposits amplifies large enterprises’ and government bodies’ incentives to run under the ECON report approach, leading ceteris paribus to a more fragile financial sector.

However, even the CMDI Proposal would not provide a comprehensive solution. DGS funding has to satisfy the ‘least cost test’ (“LCT”), according to which any DGS contribution to resolution shall not be greater than the cost of 552repaying depositors (proposed Article 11e DGSD).^[53] Hence, at banks that have a high proportion of uninsured deposits^[54] – and are thus particularly prone to runs – the available DGS funding in resolution may not always suffice to fully bridge the gap between an entity’s available loss-absorbing capacity and the 8% TLOF requirement to access the SRF. Consider an institution that has no access to public capital markets and thus refinances itself with equity and deposits, satisfying MREL exclusively with CET1-instruments. If a buyer in a sale of business transaction required a 15% markdown on the bank’s total assets and the remaining equity valued at 4% of total assets, deposits would have to be written down by 11% to meet the buyer’s demand. If 80% of these deposits were uncovered, the DGS contribution would be capped at 2.2% of total assets (assuming that the assets could also be sold for 85% of their book value in insolvency). Hence, writing down equity and receiving a DGS contribution would not meet the 8% TLOF requirement.^[55] To meet this requirement, uncovered deposits would have to be cut, potentially precipitating a run.

While the CMDI Proposal and the ECON Report are broadly aligned on the use of DGSs in resolution, the Council Report introduces stricter conditions for the use of DGS funding in resolution under its proposed Article 109 BRRD3 and excludes (i) institutions that have been identified as a liquidation entity (including in a previous resolution plan in the two years preceding resolution); (ii) institutions subject to the transition arrangement to build up MREL (with an exception in systemic crisis); (iii) banks that have breached own funds requirements during a period of 8 to 36 months before the bank is declared failing or likely to fail (“FOLTF”). Furthermore, DGS’s contributions would be possible only for a maximum of 2.5% of the bank’s TLOF and subject to additional conditions aiming to limit the duration of resolution to the minimum possible, i.e., liquidate the residual entity and (if relevant) terminate a possible bridge bank as soon as possible.

553Overall, despite some changes in the wording,^[56] the TLOF requirement is left intact. Therefore, the CMDI Proposal remains stuck halfway in implementing the legislative intention of making resolution under the CMDI framework a more attractive choice. The funding gap persists if DGS contributions are insufficient to reach the 8% TLOF requirement once an institution’s MREL resources have been exploited. This result may accord with the political intention to push resolution authorities and banks to achieve high MREL targets. However, it is heedless of the challenges such a policy approach creates for small and medium-sized banks.

5.1. Early write-down and conversion (the CoCo alternative)

The banking turmoil of March 2023 teaches a critical lesson for the CMDI review. The causes of banking stress may differ.^[65] While the US regional banks were suffering from poorly managed interest rate and concentration risks, backfiring regulatory arbitrage (held-to-maturity sovereign bonds in the banking book), and highly fragile funding structures (overreliance on uninsured deposits) without decent liquidity risk management, Credit Suisse was suffering from serious and continued weaknesses in governance, risk control, and risk management. In addition, particularly in the US, but also in Switzerland, various supervisory shortcomings amplified the problems.

558However, one overarching question in all these idiosyncratic bank failures was whether the liability side of banks‘ balance sheets allowed for sufficient loss absorption and effective recapitalization when the fragilities materialized. While the AT1 dam of preferred shares did not stem the tide at Silicon Valley Bank, Signature Bank of New York, and First Republic Bank, the contingent convertible bonds at Credit Suisse provided a sufficiently thick cushion to absorb the bank’s losses. However, significant hiccups had to be remedied with unconventional legislative interventions^[66] that prompted leading scholars in the field to ask for, among other things,^[67] improvements in the design of CoCo-bonds.^[68]

We concur with the objective of providing for sufficient recapitalization capacity that authorities can activate to stabilize a wobbly bank without triggering resolution. At a pre-resolution phase, the use of write-down and conversion of capital instruments (“WDCCI”) would, in principle, allow authorities to impose fewer losses on bank stakeholders through PSI because, at an early stage, the remaining goodwill of the bank would be higher. To be sure, more operational WDCCI would not in itself revive the prospects of the bank. Unless the bank is facing an exogenous event of extreme liquidity shortage, it is most likely vulnerable because of a failing business model. WDCCI only buys time to identify weaknesses early on and implement recovery measures to turn around the fortunes of the bank in a joint effort of the troubled institution, its banking and market supervisiors, and resolution authorities. If markets and depositors are not confident that the essential recovery will succeed, an early WDCCI may only send a counterproductive signal that accelerates bank runs.

In any case, we do not think prudential regulators should try to make WDCCI operational by continuing the thus far unsuccessful CoCo experiment. Requiring a layer of going concern loss absorbing capacity that the supervisor can write down at an early stage of an ensuing crisis can be achieved consistently with minor tweaks of the current European regulatory framework. BRRD art. 59(1) allows the writing down not only of own funds but also eligible liabilities, as defined in BRRD art. 2(71a), once an institution reaches the PONV. Therefore, in principle, MREL instruments and resolution authorities’ 559write-down-and-conversion-powers provide precisely the loss absorption capacity in a going concern scenario that Perotti and Martino seek to establish with CoCos that empower the regulator to trigger the recovery. The only adjustment required in the regulatory framework is an adaptation of the PONV take-up criterion in BRRD art. 59(1). Instead of requiring a failing or likely to fail determination,^[69] the WDCCI tool could become available once indicators, suggested in the literature, provide growing evidence that the bank is in severe trouble.^[70] A future CMDI reform could implement such an incremental change without much effort and stay on the European resolution trajectory that puts MREL at center stage.

5.2. Effective auctions in transfer strategies

The second additional element future CMDI reforms should consider is the careful pre-structuring of auctions in transfer strategies. Making industry resolution funding more readily available remedies a looming lack of sufficient loss absorbing capacity in the bank. It provides the essential precondition for achieving the overarching policy objective. Sufficient resolution funding allows to devise optimal resolution schemes that minimize the social cost of resolution. In the context of small and medium-sized retail banks, it increases the chance of success of transfer-based resolution strategies (compared to an open-bank bail-in strategy which would put additional pressure on the bank for internal loss-absorption and recapitalization). However, research shows that ad hoc fire sales of failed institutions in resolution may lead to windfall profits for acquirers and their shareholders because resolution authorities leave money on the table to the detriment of all other stakeholders.^[71] Therefore, a sound policy that seeks to improve the regulatory environment for transfer strategies in bank resolution, should also optimise the framework underpinning the transfering transactions.

Generally speaking, economic theory favors using auctions instead of negotiations when authorities want to sell a failed bank to a healthy one. The basic intuition is that when bidders have private information about the value of the object on sale, as is likely the case when acquiring a failed bank, using an auc560tion compels bidders to reveal some of this information; even simply leaving an English auction, in which the price rises until one of the bidders drops out, reveals some of the private information. Such an information revelation leads to more efficient outcomes because acquirers earn a lower information rent – akin to a monopoly rent – in an auction.^[72] The information revelation benefits the uninformed seller (e. g., the resolution authority acting on behalf of the failed banks’ equity holders, unsecured creditors, and taxpayers) and prevents purely redistributive transactions.

However, to reap the potential welfare gains from an auction, banking supervisors and resolution authorities need to adequately prepare for a bank resolution that may come unexpectedly and suddenly due to a bank run. Assets that are carved out because of valuation uncertainties need to be identified. Due diligence materials and information for potential bidders must be prepared early – long before the auction starts and even before the decision to sell the bank in an auction is ultimately taken. Resolution planning should enable the authorities to initiate a professionally managed sale process and due diligence. Resolution authorities also need to prepare the “Best Alternative to a Negotiated Agreement” (BATNA) carefully to avoid hold-up scenarios.The essential preparatory work needs to be a part of resolution planning, a core task of resolution authorities, including the SRB.

Quite importantly, resolution authorities also need to prepare for transfers of non-significant banks. For example, SVB was not designated as a global systemically important (G-SII) or other systemically important institution (O-SII) by the relevant authorities in 2022. SVB was large but not considered to be complex enough and interconnected enough to warrant significant institution status. If a similar bank were to fail in Europe, the national laws would apply because the SRB de facto only assumes jurisdiction over significant institutions due to the PIA (see above 4.2.3). There is little doubt that regular bankruptcy would have created massive uncertainty and systemic risk and triggered the need for a full-blown government bail-out. This observation provides another argument for widening the scope of the CMDI framework and making it fit for purpose, particularly for dealing with smaller banks in sale of business and other transfer strategies.