The Single Resolution Fund should be sufficient to deal with almost any crisis scenario imaginable. However, the funding for the Single Resolution Fund will be built up only gradually over the coming decade. There is thus a legitimate concern that the Single Resolution Fund will not be sufficient to deal with a major crisis during the transition.
How much does the Single Resolution Fund need?
To address this concern, the European Parliament asked us to perform an exercise that is straightforward in principle, but fiendishly difficult in reality – to calculate how much funding would have been needed from the Single Resolution Fund during the last banking crisis. We thus took all the balance sheet data from all Eurozone banks that had received state aid since 2007 and assumed that the new rules on bail-in – the Single Resolution Mechanism and bank recovery and resolution Directive – were already in force.1
We then calculated how much funding would have been required from the Single Resolution Fund to keep these banks afloat. The result was that a total of about €72 billion would have been sufficient in a central case. Under a more stringent assumption about the capital requirements for the resolved banks, the sum might go up to €102 billion, or down to €54 billion under more optimistic assumptions (De Groen and Gros 2015).
How much can the Single Resolution Fund get?
The central case of €72 billion in funding requirements from the Single Resolution Fund should be compared to its target size of 1% of covered deposits, or approximately €55 billion. Moreover, the SRF would be able to impose ex-post contributions of 3/8 of the total, thereby increasing the total funding at its disposal to over €75 billion.
The Single Resolution Fund would thus have had enough funding to deal with the biggest banking crisis in European post-war history. But how is this surprising result possible?
The number of banks that were involved in ‘interventions’ during 2007-2014 was relatively low – only 72 banks received state aid in the form of capital injections, a fraction of all banks. Of these, data are available for 62 banks with assets of around €14 trillion, equivalent to approximately 45% of the total Eurozone banking assets.2 This was certainly a system crisis. Most of the banks in trouble would be classified as significant under the Single Supervisory Mechanism/ECB criteria.3
The problems were unevenly distributed: banks needing support accounted for more than 70% of the banking assets in four countries – Belgium (3 banks), France (6), Greece (12) and Portugal (7). But both Germany (10) and Spain (13) many banks received state aid, but they accounted for less than 35% of the national banking system.
Why is the Single Resolution Fund contribution only about a quarter of the total losses?
The total losses incurred by the banks needing support amounted to about €313 billion. This figure looms large relative to the size of the Single Resolution Fund, thus deepening the mystery. But the same figure also appears small relative to the total balance sheet of these banks – less than 3% of total assets.
There are two solutions to the mystery. First of all, there is bail-in. The Single Resolution Fund can intervene only after a bail-in of 8% (of liabilities) has taken place. Writing down 8% of liabilities provides a first protection for the Single Resolution Fund. Second, there is a limit to the funding the Single Resolution Fund is authorised to provide – it can extend only up to 5% (of liabilities including own funds).
One finds that these two elements would have been important at different stages of the crisis. We find that for the largest banks (those in difficulties during the first leg of the crisis), the 8% bail-in took care of most of the losses, leaving little need for funding from the Single Resolution Fund. During the second leg of the crisis, the 8% bail-in was usually not sufficient, but here the Single Resolution Fund would often have been limited by the 5% ceiling. In these cases additional bail-ins might have been needed. There is a technical point here – the reduction in funding needs from the Single Resolution Fund resulting from bail-in depends on the ratio of risk-weighted assets to total liabilities (which equal assets) because the recapitalisation need is measured as the percentage of the risk weighted assets whereas the bail-in is based on total assets. For a ‘classic’ banking model where all the lending is to the corporate sector, risk weighted assets would be equal to total assets. In this case, losses might be absorbed by a bail-in, but the bank would still have to be recapitalised to 8% of risk-weighted assets. By contrast, for a bank with other business, which has a lower risk weight and a low ratio of risk-weighted assets to total assets, the 8% of total assets bail-in would be more than sufficient to cover the recapitalisation need. A corollary of this technical point is that bail-in reduces the funding needs from Single Resolution Fund mainly for large banks.
The result of all this can be seen in Figure 1 which shows a Lorentz curve for the funding the Single Resolution Fund would have had to provide for the 62 banks with data available mentioned above. If one orders banks by the hypothetical contribution they would have received from the Single Resolution Fund the first 40 banks would have needed nothing and the biggest handful of cases would have absorbed about one half of the total funding needs from the Single Resolution Fund. But the first 40 banks without financing need from the Single Resolution Fund would have amounted to 80% of the assets of the banks in trouble and 40% of the total losses. The current rules on bail-in would thus have ensured that a large part of the losses would have been borne by investors.
Source: Own elaboration on data from De Groen and Gros (2015).
Another result of the interplay between risk weighted assets and total assets combined with the ceiling of 5% (of liabilities) for the Single Resolution Fund is that the link between losses and funding needs is not as tight as one might think. Figure 2 below shows a scatter plot with the losses on the horizontal and the hypothetical contribution of the Single Resolution Fund on the vertical axis. A best fit line can explain only about one half of the variability in the Single Resolution Fund contribution and the four big cases in the oval on the right illustrate that for losses around €25 billion the contribution of the Single Resolution Fund could have been either zero (Fortis, with a low ratio risk-weighted assets/total assets) or €15 billion (Bankia). Moreover, almost all cases lie below 50%, meaning that only in very few cases would the Single Resolution Fund have covered more than one half of the losses.
Source: Own elaboration on data from De Groen and Gros (2015).
But the funding at the disposal of the Single Resolution Fund during the transition would be much lower – would it not surely be insufficient under a replay of the Global Crisis?
This type of question is based on a distorted view of banking crises. The eruption of the problem can be dated to a single point in time. But the losses arise over a longer period of time during which the crisis affects the economy. The banking crisis in Europe that started in 2007 is no exception – the losses arose over a period of eight years, which happens to coincide with the time period needed to fully ‘load’ the Single Resolution Fund. It is clear that a Single Resolution Fund in the building-up phase would not have been sufficient if all the losses that arose over the past eight years were to materialise again in a single year. But this is extremely unlikely. The central scenario would thus be that the gradual build-up of the Single Resolution Fund should be sufficient to deal with a major crisis, even one of similar proportions as the last one. Some limited bridge funding might be needed in case the next crisis is very much ‘front-loaded’, but our paper shows that the necessary mechanisms exist, or could be mobilised on short notice (see also Zavvos and Kaltsouni 2015).
The need to ensure the credibility of bail-in
The upshot of this research is therefore that applying the new rules on bail-in will be crucial when the next crisis arrives. But bailing-in on a large scale will be possible only if this does not lead to contagion. Care should thus be taken to maintain the credibility of the bail-in procedures during good times (Huertas and Nieto 2014) and to limit cross-holdings within the banking sector of the capital instruments potentially subject to bail-in (Krahnen 2014).
If these two objectives can be achieved, the Single Resolution Fund should be sufficient.
That the Single Resolution Fund might have sufficient funding to formally deal another crisis like the last one does not mean that all is well. No resolution fund can be expected to deal alone with a major systemic crisis. If another large-scale crisis erupts, which becomes systemic, at the Eurozone level, the Single Resolution Fund will need of course a fiscal backstop. The absence of such a backstop has often been taken as meaning that the Single Resolution Fund is useless (Schoenmaker 2014). But member states have now committed to develop a common backstop by the time the national compartments in the Single Resolution Fund have been fully mutualised.4 Once this has done, the Banking Union should have two lines of defence which should be sufficient to protect systemic stability against even larger crises.
De Groen, W P and D Gros (2015), “Estimating the Bridge Financing Needs of the Single Resolution Fund: How expensive is it to resolve a bank?”, CEPS Special Report No. 122, CEPS, Brussels, November.
Gros, D (2015), “Completing the Banking Union: Deposit Insurance”, CEPS Policy Brief No. 335, CEPS, Brussels, 3 December.
Huertas, T and M J Nieto (2014), “How much is enough? The case of the Resolution Fund in Europe”, VoxEU, London, 18 March.
Krahnen, J P (2014), “Implementing bail-in properly”, Policy Letter No. 35, SAFE, Frankfurt, 13 November.
Micossi, S (2014), “Bail-in rules in EU banking union and financial stability”, VoxEU, 5 June.
Schoenmaker, D (2014), “On the need for a fiscal backstop to the banking system”, DSF Policy Paper No. 44, Duisenberg school of finance.
Zavvos, G S and S Kaltsouni (2015), “The Single Resolution Mechanism in the European Banking Union: Legal Foundation, Governance Structure and Financing”, in M Haentjens and B Wessels (eds.), Research Handbook on Crisis Management in the Banking Sector, Cheltenham: Edward Elgar Publishing Ltd.
1 For a description of the new rules see Micossi (2014).
2 The total banking assets (i.e. Monetary Financial Institutions excluding Eurozone central banks) were on average €31.5 trillion (varying between €26.4 and €34.9 trillion) during the period from January 2007 to December 2014, see ECB Statistical Data Warehouse (2015, available at https://sdw.ecb.europa.eu/quickview.do?SERIES_KEY=117.BSI.M.U2.N.A.T00.A.1.Z5.0000.Z01.E.
3 Banks are considered significant if the assets are over €30 billion or over €5 billion and 20% of GDP or if the bank is one of the three largest banks in a country or has significant cross-border activities (ECB 2014).
4 See the Press Release of 8 December, available at http://www.consilium.europa.eu/en/press/press-releases/2015/12/08-statem....