Completing the Banking Union

Daniel Gros

12 February 2016



The consensus narrative (Baldwin et al. 2015) concluded that “Tthe “EZ cCrisis was a ‘sudden stop’ crisis”. The government debt crisis came later and was a consequence of the sudden stop, as can be seen from the fact that “[Aa]part from Greece, the nations that ended up with bailouts were not those with the highest debt-to-GDP ratios” (Baldwin et al. 2015).

What made the crisis so costly in terms of output and employment was not the higher risk premia on government debt per se, but the tight link between governments and bank finances, combined with the predominance of bank financing throughout Europe. The ‘doom loop’ acted as a crisis amplifier whereby weak governments increased the cost of financing for banks and the difficulties of banks then depressed the economy, which then weakened both governments and banks even further.

EZ leaders finally recognised this problem when they decided in the summer of 2013 to create a system of banking supervision and a common institution to restructure banks in difficulties. This combination has been dubbed ‘Banking Union’, but, while important and helpful, what has been done so far is not sufficient to prevent the recurrence of crisis. The Banking Union should be completed and complemented with other measures to weaken the link between national governments and national banking systems.

Two issues in particular need to be tackled:

  1. Banks hold too much debt of their own sovereign. A sovereign default would bankrupt the banks in the country (e.g., Greece).
  2. Deposit insurance has been left in national hands with only a national back-up. But if there is a national crisis the government might not be able to provide a credible backstop for deposits (e.g., Ireland).

The two issues need to be tackled in tandem for economic and political reasons, but the required solutions are quite different in nature – diversification of sovereign debt holdings can be enforced by a low-level change in some obscure details of banking regulation (and maybe even by the ECB on its own). A common deposit insurance requires the creation of a new institution.

Diversifying sovereign risk

The first problem (excessive holdings of own government debt) seems relatively straightforward, but any sensible solution encounters strong political resistance.

The instability created by large holdings of sovereign debt by banks has been widely recognised – most recently see the Advisory Scientific Committee (ASC) of the European Systemic Risk Board (ESRB). The problem is not how much government debt banks hold, but the concentration. In many countries banks hold debt of their own sovereign equivalent to more than 200% of their capital. This means that any sovereign restructuring would bankrupt the banking system. Formal sovereign defaults are rare, but even a sharp increase in the risk premium, which would depress the market price of long- term debt, would bring the banks into difficulties.

An example can illustrate the importance of this feedback mechanism – risk premia of 5 percentage points (as reached by Italy or Spain at the peak of the crisis) can reduce the market value of long- term government debt by almost 40%. Given that Italian banks held about 250% of their capital in Italian government debt, this risk premium would wipe out their capital. (Banks are not required to mark their holdings of government debt to market. The official capital ratios were thus not affected by the risk premium. But the market saw through the official balance sheets and marked down Italian banks).

The situation of Italian banks would have been quite different if they had held a diversified portfolio of EZ government debt. In this case the fall in value of peripheral government debt would have been compensated by an increase in the value of core-country debt (the price of German debt went up as capital fled back from the periphery). It would not have mattered whether the banks had held 250% of their capital or even more in government debt – with appropriate diversification, the high and variable risk premia during the euro crisis would have had little differential impact on the solidity of banks in different countries.

The key problem is thus the concentration of sovereign risk on bank balance sheets, not how much government debt banks hold (see ASC and Gros 2013a).

How could this diversification be achieved? Banking regulation already embodies the general principle that banks should not be exposed to any one debtor by an amount greater than a quarter of their capital. The purpose of this general rule is clear – the insolvency of any one debtor should not put the entire capital of the bank in jeopardy. Unfortunately, this rule is not applied to sovereigns. Gros (and others) have thus proposed simply applying this general ‘large- exposure’ rule to government debt as well. Others have proposed the creation of a synthetic EZ government debt by securitising a basket of EZ government debt (Euronomics).(see De Groen (2015).

Action along these lines is thus needed. Until now it has proven impossible to get governments to agree on some diversification rules for bank holdings of government debt because every government likes to have its own banks as captive customers for its own debt. Moreover, banks (and governments) in the periphery would lose the income they earn by holding the higher- yielding bonds of their own government. This is important for the banks – even if the difference between national government debt and the EZ average is only 1-2 percentage points, holdings of national government debt (instead of a diversified EZ portfolio) would mean a lower return on equity of between 2.5 to 5 percentage points. Given that the return on equity is in single digits in any event for many banks today, this represents an important loss for shareholders. But dealing with the second issue, deposit insurance, could provide a way to overcome this resistance.

Deposit insurance

The report of the five presidents argued that a banking union should not only encompass common supervision and a common restructuring mechanism, but also needed a common defence of retail deposits to prevent bank runs. This is the case in the US. Gros and Schoenmaker (2012) argue that one should work backwards from the end game – when a bank is failing and depositors need to feel protected – to restructuring as a way to avoid outright failure, and supervision to ensure that banks do not take undue risks.

However, this does not imply that resolution and deposit insurance should always be located in the same institution with the Federal Deposit Insurance Corporation (FDIC).

In reality, resolution and deposit insurance are of a quite different nature; and the rationale for fully centralising deposit insurance is much weaker than that for resolution. The purpose of bank resolution is to avoid insolvency, with all the costs and contagion effects it might generate. Resolution thus aims to ensure continuity of those main functions of a bank that are deemed to be of systemic importance. Public funding is needed only to the extent that no private-sector solution can be organised on short notice. The purpose of a resolution fund is to finance investment in a new bank (to be carved out of the failing one) – not to give money away. A well-run resolution fund should thus be profitable. By contrast, a deposit insurance fund can only make losses, as it is used when a bank has failed and the losses are so large that depositors cannot get their money back. In short, a resolution fund invests in the future, whereas a deposit insurance fund pays for losses from the past.

Several European banks have very large balance sheets, close to one year’s worth of national output. But large banks active in several member states would represent a problem for any national resolution fund – though much less for national deposit insurance funds, since any one of them would only have to make good on the losses suffered by resident depositors (provided the retail operations are organised via subsidiaries, rather than branches). By contrast, resolution funding is needed for the entire group. This is another reason why a single resolution fund is needed for large, internationally-active banks, but not necessarily a single deposit insurance fund.

Moreover, the direct benefits of insuring depositors are quite local, as cross-border retail deposits remain rare. It thus makes sense to keep the cost local as well. All this implies that the argument for centralising deposit insurance is much weaker than for bank resolution.

The one public function that national deposit insurance funds are not well equipped to perform is that of maintaining the confidence of depositors when the entire banking system of a country is under stress. In a systemic (at the national level) banking crisis, the accumulated funds in the national deposit insurance scheme are likely to be insufficient. But the government of the country in question is likely to be under pressure as well, which implies that the national fiscal backstop is likely to be weak when it is needed most.

This implies that what is needed to complete the Banking Union is a system to ensure the stability of the national deposit insurance system – in other words re-insurance (for large, systemic shocks at the national level). The Commission proposed a variant of this re-insurance approach in December of 2015.

The re-insurance approach: A schematic presentation

Under the pure re-insurance approach, national deposit guarantee systems (DGSs) would continue to function as before, but each one would be forced to take out insurance coverage against large shocks. This (re-)insurance could be managed under the European Deposit Insurance System (EDIS), as it is called in the Commission’s proposal. The funding for the re-insurance could come from a common fund (like the Deposit Insurance Fund, DIF, also proposed by the Commission).

The funding for the DIF in turn would come from the national DGSs, which would have to transmit part of the fees they are levying on individual banks to the DIF at the European level). Schematically there would thus be two tiers of deposit insurance –the national DGSs in relationship to ‘their’ banks, and the European re-insurer in relationship to the national DGSs.

Figure 1. The two-tier approach to deposit insurance: Re-insurance

This two-tier system would react differently to the failure of a small bank than to a systemic problem at the national level. This is illustrated in Figures 2 and 3.

Figure 2. The case of a single bank failure

Figure 3. Systemic crisis at the national level

The European re-insurer (i.e., EDIS with its fund, the DIF) would thus intervene only if so many banks fail in any given country such that the national DGS would be overwhelmed.1

It is clear that the resources of any normal deposit insurance will always be too small in the event of a systemic crisis. This also applies for the EZ as whole. If there is a systemic crisis at the EZ level (as opposed to a national crisis), any European insurance scheme would need a fiscal backstop. This applies to the re-insurance approach as well as the case in which there would be one single European deposit insurer. Figure 4 depicts schematically the case of an EZ-wide crisis, assuming that the ESM would constitute the fiscal backstop.

Figure 4. The case of an EZ-wide systemic crisis and the need for a fiscal backstop

Macro risk-pricing and alternative private-sector solutions

One key aspect of the reinsurance approach is that it is a macroeconomic function. Its main concern will not be the risk parameters of each individual bank in each country, but rather the systemic risk that arises from developments at the macroeconomic level (such as rising housing prices, increasing leverage in the corporate sector, etc.). In principle, this expertise is already available in the European Systemic Risk Board (ESRB). It would thus be important to find an institutional solution under which this expertise can be used. Moreover, procedures to prevent systemic problems already exist; for example, the Macroeconomic Imbalances Procedure (MIP) and the Growth and Stability Pact (GSP). The macroeconomic aspect will remain a predominantly national responsibility for a long time, driven by national fiscal policy, wage developments, the fiscal treatment of housing, etc. These are the risks that would be priced under the re-insurance approach.

Pricing macroeconomic risk will always remain imperfect, but some risk pricing is better than none. Gros (2015) illustrates how one could introduce an element of ‘experience rating’ to minimise the scope for cross-country transfers if that were to prove politically needed.

One could argue that the re-insurance of (national) DGSs could also be achieved through a private-sector solution, in a similar way to how re-insurers provide cover against large risk to insurance companies. However, a private sector solution is unlikely to work in this area because the risk is primarily of a macroeconomic nature.

Moreover, the capital base of even the largest re-insurance companies is only around 25 billion euro, not sufficient to provide a back-up even if only a medium-sized country experienced a systemic crisis. By contrast, the DIF would be able to count on the back-up of the constituent national DGS whose combined funding potential will be closer to 80 billion euro.

Moreover, the largest reinsurance companies are European, and would thus likely be in a difficult situation themselves in the case of a systemic crisis affecting the entire EZ. This is what happened to the so-called monoline insurance companies in the US, whose primary business had been to insure the debt of municipalities, but which had insured bonds based on sub-prime mortgages. Those most involved in this business went bankrupt when the US housing market tanked in 2008/9.


Nation states are usually able to deal with small shocks themselves, but they need support when the shock is so large that access to the capital market is impaired (Gros 2014). In completing the Banking Union, one should heed the lessons from the economics of insurance and provide protection against large, systemic shocks. Nation states could remain responsible for deposit insurance for the occasional individual bank failure. But a common fund is needed to provide re-insurance when the shock is so large that it would overwhelm national resources.

A similar principle should be applied to banks – they should be able to survive the failure of any one of their debtors, including the failure of their own government. But this means that banks should not be allowed to lend more than a fraction of their capital to their own government.

Peripheral governments still resist mandatory diversification of sovereign risk for banks, but would like more risk-sharing through a common deposit insurance. Germany has taken the opposite position. The package of sovereign-risk diversification by banks and risk-sharing through re-insurance of deposit insurance could represent a political compromise that makes economic sense as sovereign risk diversification by banks would lower the danger of systemic crisis caused by imprudent fiscal policy.


Baldwin, R et al. (2015) “Rebooting the Eurozone: Step 1 – Agreeing a crisis narrative”,, 20 November.

De Groen, W P (2015),The ECB’s QE: Time to break the doom loop between banks and their governments”, CEPS Policy Brief No 328, CEPS, Brussels, March.

Eichengreen, B and C Wyplosz (2016), "Minimal Conditions for the Survival of the Euro", Intereconomics 51(1): 24-28, January/February.

Gros, D (2013a) “Banking Union with a sovereign virus: The self-serving regulatory treatment of sovereign debt in the Eurozone”, CEPS Policy Brief No. 289, CEPS, Brussels, 27 March.

Gros, D (2013b) “Principles of a Two-Tier European Deposit (Re-)Insurance System”, CEPS Policy Brief No. 287, CEPS, Brussels, 17 April.

Gros, D and D. Schoenmaker (2012) “The case for euro deposit insurance”,, 24 September.

Gros, D (2014) “A fiscal shock absorber for the Eurozone? Lessons from the economics of insurance”,, 19 March.

Gros, D (2015) “Completing the Banking Union: Deposit insurance” CEPS Policy Brief No. 335, December 2015.


1 The one risk that cannot be properly priced and prevented is that of re-denomination, or rather exit from the euro. It is clear that no common deposit insurance could be asked to fully pay out €100,000 to all depositors in a country where the government has decided to re-introduce a national currency. The ‘Grexit’ problem cannot be solved either by fully centralised deposit insurance or via the re-insurance approach. The only solution one can imagine is that any country that leaves the euro will have to rely on its national DGS, which could claim from EDIS and the DIF only the amount of premia it had paid in previously.The European Re-insurer should thus be backstopped by the ESM. This would be a natural evolution of its role since it was set up to backstop countries and can already now lend directly for banking resolution.



Topics:  EU institutions EU policies Europe's nations and regions

Tags:  reboot, EZ, eurozone, deposit insurance, sovereign debt, politics, sovereign risk, re-insurance, systemic risk, risk pricing

Director of the Centre for European Policy Studies, Brussels