This blog post, written together with Isabel Schnabel, was published by PIIE and Bruegel.
In the wake of the European financial and sovereign debt crisis, the euro area embarked in 2012 on establishing a banking union. Its aim was to elevate parts of banking sector policy from the national to the European level, particularly bank supervision and resolution. Successive EU-level reports, including the Four Presidents’ Report of 2012 and the Five Presidents’ Report of 2015, have highlighted a European Deposit Insurance Scheme (EDIS) as a necessary component of banking union. In 2015, the European Commission published a legislative proposal (hereinafter EC 2015) to set up a fully integrated, country-blind deposit insurance system by 2024.
A number of different proposals have arisen in this area. A recommendation by Daniel Gros (Gros 2015), director of the Centre for European Policy Studies, would retain a permanent autonomous role for the existing national deposit insurance schemes in a re-insurance system. Meanwhile, none of these options has met sufficient consensus among euro-area countries, producing a deadlock in the policy discussion, with no apparent progress in the legislative discussion of the EC 2015 proposal.[1]
In a Franco-German report in January, we proposed, jointly with our coauthors, to end the deadlock with an EDIS design that is institutionally integrated but financed in a way that is differentiated across countries. Our recommendations are outlined here and compared with EC 2015 and Gros 2015.[2]
EDIS as part of a broader policy package
Our EDIS proposal is tied to the introduction of sovereign concentration charges and tighter treatment of nonperforming loans (NPLs). It calls for “the coordinated introduction of sovereign concentration charges for banks and a common deposit insurance.” This connection is necessary to prevent the system from being “abused by governments that can force or nudge domestic banks to grant them preferential credit conditions by using their access to deposit funding.” By contrast, EC 2015 does not include any proposals on the regulatory treatment of sovereign exposures. In his own proposal, Daniel Gros observes that he “has not addressed the issue of the large holdings of banks of the debt of their own government” while hinting at “strict diversification limits,” which echo our recommendation of sovereign concentration charges. As for NPLs, our proposal calls for the European Central Bank (ECB) to foster better accounting for these loans on banks’ balance sheets and reduce their aggregate volume, including for smaller banks principally supervised by national authorities. The report calls for “full implementation of a uniform regime for NPL provisioning covering both legacy and new NPLs.”
A fully integrated institutional setting
A major difference between our proposal and Gros 2015 is that in our proposal, national deposit insurance schemes disappear after a transition period, replaced by a system with “a single authority at the European level,” which we suggest should be the Single Resolution Board (SRB) with due adjustments, and the existing “separate national deposit insurance institutions would be phased out.” In the event that the EDIS scheme would need to make direct payouts to individuals, it would rely on national authorities, e.g. the national resolution authority, but this arrangement would be purely for implementation purposes and the European authority would make all the significant decisions.
As a complement to our EDIS proposal, new EU legislation should eliminate the widespread practice of geographical ring-fencing of capital and liquidity by national authorities, which hinders the emergence of cross-border banking groups in the euro area. Since the protection of national deposit insurance schemes is the main reason cited to justify such practices, their phasing-out would make this reform possible and increase the potential for cross-border banking acquisitions, contributing to breaking the bank-sovereign vicious circle.
Our proposal envisions country-blind protection of insured deposits. After the end of the transition period, Cypriot or Greek insured deposits would be protected identically to German or Finnish ones, without any qualification or conditionality. This is essential to establish trust. We fear that the deposit re-insurance concept in Gros 2015, which perpetuates autonomous policy decision-making on deposit insurance by national authorities, could lead to uncertainty as to the automaticity of payouts, especially in cases where a country’s politics or policies are controversial in the rest of the euro area—a possibility that was illustrated in March 2013 in Cyprus. Thus, the operation of the system from the insured depositor’s perspective must be fully insulated from national political vagaries, as it is in our proposal (and in EC 2015).
A funding framework that recognizes national differences
The system would not be country-blind in terms of its funding mechanism, a key difference with EC 2015 but a similarity to Gros 2015. Our analysis is that despite the progress towards the completion of banking union, differences in national environments that affect banks’ business models will persist in terms of taxation, corporate and individual insolvency frameworks, housing finance, pension finance, and macroprudential policies. (All such policies are comparatively harmonized in the United States, in which there is correspondingly no need for state-level differentiation within the federal deposit insurance system.) This justifies differentiated assessments of bank risks and, correspondingly, deposit insurance fees across euro-area countries. As in most existing deposit insurance schemes and in EC 2015, there would also be fee differentiation at the level of individual banks to account for diverse risk profiles.
We have received objections that this would make EDIS different from its sibling, the Single Resolution Fund (SRF), which is already in place and will be fully mutualized by 2024. But since the European Union has chosen to distinguish resolution funding from deposit insurance (a distinction that does not exist in some other jurisdictions, e.g. the United States), and political sensitivities are evidently different about the two areas, there is no need for identical policy responses to the challenges tackled by EDIS and the SRF respectively. We do suggest, however, that both be managed by the same EU authority, namely the SRB.
Our proposal calls for two separate modalities for such country-level differentiation:
- First, a fee component depending on country-specific risks, such as the quality of the country’s legal framework for creditor protection, “based on structural indicators of creditor rights, such as the effectiveness of insolvency and foreclosure processes.” With harmonization of legal frameworks, fees would also converge. These indicators would have to be designed and measured at the European level, possibly by the SRB itself, or by an adhoc independent advisory body reporting to the SRB. The aim is to establish good structural credit policies in foreclosure and other areas, underlined by developments in Greece and Cyprus.
- Second, a structure requiring the cost of payouts associated with a bank failure to fall on banks in the same country in case of limited idiosyncratic shocks but to be mutualized in case of a larger crisis. The EDIS system would contain national compartments, akin to the ones that exist in the SRF during the current transition but which (unlike at the SRF) would be kept in the longer run. Two possible designs are a mutualized compartment alongside the national ones, similar to the transition-state SRF (“option1”), or a joint payout by all compartments in case one of them is depleted (“option 2”). In option 1, the ex-ante deposit insurance fee (i.e. in the absence of any bank failure triggering the insurance) would be split between the national and mutualized compartments, with a fixed allocation key set in the EDIS legislation that would apply identically to all member states. In both options, if a national compartment is depleted by payouts, there will be an ex-post fee paid only by banks domiciled in the corresponding country, until it is replenished. If the European compartment is itself depleted (option 1), or the joint capacity of all compartments to step in is extinguished (option 2), the ESM would step in as a backstop (as with the SRF)—which would not be a loss to the ESM, as it would be reimbursed by banks thanks to appropriate ex-post fees. The economics of this system are similar to those in Gros 2015, even though the institutional arrangements are fundamentally different. This difference is why we refer in the report to “the spirit of a re-insurance system,” with a first loss at the country level. The legal arrangement would be one of direct insurance, not re-insurance. Thus, unlike in Gros 2015, the national compartments we propose would be very different from the existing national deposit insurance schemes, which would no longer exist in the steady state as explained above.
Country-level differentiation would create sound incentives. Banks could choose to operate in a given member state either through a locally capitalized subsidiary or through a branch of an entity established in another (home) country. If it is a subsidiary, the deposit insurance (or in our proposal, the national compartment) would be that of the host country; if a branch, that of the home country. If a country has frequent bank failures that trigger deposit insurance payouts, or if its structural credit policies are unsound, the higher resulting deposit insurance fees may encourage banks to serve clients in that country through branches rather than subsidiaries. Such a country’s national compartment would correspondingly decrease in size, due to the reduction in the country’s total covered deposits, while other countries’ compartments would grow. This is how it should be.
Importantly, our proposal of country-level differentiation would have nothing to do with a country’s sovereign credit, and would therefore not contribute to the bank-sovereign vicious circle. The proposal would also create incentives for countries to reduce competitive distortions, such as those resulting from national housing finance and pension finance frameworks, insolvency law, or bank taxation.
Our EDIS proposal can accommodate the present diversity of banking structures in the euro area. In 14 of the area’s 19 countries representing together more than half of total covered deposits,[3] there is currently one single national deposit insurance scheme. But in five other countries, there are two or more schemes serving different segments of the national banking sector.[4] We suggest that these “could be treated as separate compartments, on a case-by-case basis under general criteria to be set to deter abuses.” These criteria could also create incentives to merge such schemes on a cross-border basis, e.g. to protect deposits in all the cooperative banks of several euro-area countries. Voluntary top-up regimes could also continue to exist, either on a national or a cross-border basis, with the understanding that our EDIS proposal (including any separate compartments for specific banking structures) only covers insured deposits up to the common limit currently set at €100,000.
Conversely, we advise against carving out small banks from the scope of EDIS (as is currently the case with the SRF), even those that participate in institutional protection schemes. Such small banks, or Less Significant Institutions (LSIs) as they are called in the jargon of the banking union, contribute at least as much as SIs to the bank-sovereign vicious circle, because their activities are typically less diversified across borders. We recommend that “an asset quality review of all LSIs, directly involving both the ECB and the European authority entrusted with EDIS [i.e. the SRB], would be performed as part of the implementation of the EDIS legislation, as was the case for SIs in 2014.” This step is important to ensure that the risk-sharing involved in EDIS is not abused to rescue banks that would be unsound from the start.
A winning option for all euro-area member states
The combination of country-blindness for the insured depositors but not for the contributing banks has not previously been proposed, as far as we know. But it would effectively serve to reconcile the twin objectives of risk-sharing and market discipline, the leitmotif of the entire report.
Our proposal is realistic and ready to be adopted in 2018. It would make retail bank runs much less likely in sovereign stress scenarios, and would thus reduce the redenomination risk exposed during the euro-area crisis. All euro-area countries would benefit, especially those with high debt or perceptions of credit risk, vastly offsetting any perceived downsides of introducing the sovereign concentration charges that are an indispensable complement to EDIS.
There is of course a need for further technical elaboration, not least on the transitional arrangements. The transition should not be too short. Somewhere between five and ten years is probably apt. For practical reasons, the legislative process probably has to wait until after the European Parliament election of 2019, but the Council could give a firm mandate, including a set duration for the transition, after a negotiation that need not last more than a few months. As our report puts it, “the key decisions in this area can and should all be taken in 2018.”
[1] A more recent Communication of the European Commission on banking union (October 2017) attempted to break this deadlock, but proposes only transitional arrangements.
[2] While this blog post was jointly written by the two of us in personal capacity, it was reviewed by the entire group of the report’s coauthors and there were no disagreements within that group about its content.
[3] Based on European Banking Authority (EBA) data as of end-2016. The 14 countries are Belgium, Estonia, Finland, France, Greece, Ireland, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Slovakia, Slovenia, and Spain.
[4] Specifically, and based on the same EBA data as of end-2016 as in the previous footnote, there are five separate deposit insurance schemes in Austria (respectively for savings banks, Raiffeisen banks, Volksbanks, regional mortgage banks, and commercial banks; there is an ongoing process to merge the mandatory components of these into one single national system); two in Cyprus (for cooperative banks and other banks); four in Germany (for cooperative banks, savings banks, public-sector banks, and commercial banks); two in Italy (for credit cooperative banks and other banks); and two in Portugal (for mutual agricultural banks and other banks). As per the EBA dataset, every single one of these dedicated insurance schemes has more covered deposits than the national deposit insurance scheme of Estonia.