VoxEU Column EU institutions EU policies Europe's nations and regions

How to design a banking union that limits systemic risk in the Eurozone

The Eurozone is attempting to resolve the problem of systemic risk within its ailing banking sector. This paper argues that while banking union within the Eurozone is a very real solution to this issue, it must be orchestrated correctly in order to succeed.

A key objective of bank regulation and supervision is to reduce systemic risk, that is, the risk that a large number of banks experience stress at the same time. In such situations, lending in the economy is likely to be impeded and a credit crunch may occur, leading to a recession and widespread defaults. Besides being costly, the resolution of systemic crises is also relatively burdensome. It is hence of paramount importance to have a financial structure in place that keeps the risk of systemic crises at bay.

A banking union has the potential to reduce systemic risk in the Eurozone. However, I argue that it also poses significant new challenges for the management of such risk. In this column I explain how they can be tackled. In particular, I identify four elements that a successful banking union would need to incorporate:

  • A banking union should not just lead to a simple pooling of risks, such as by centralising national deposit insurance systems. This runs the risk of making systemic crises more likely. A two-tiered structure of national and European systems is desirable, ideally with an additional European backstop in the case of systemic events.
  • When harmonising regulation, European supervisors should not fall into the danger of encouraging more similar financial systems across countries. In order to mitigate the risk of systemic crises, we need a diversity of approaches to financial intermediation in the Eurozone.
  • A European supervisor has to avoid the build-up of systemic imbalances. For this a truly systemic perspective needs to be taken since even if individual countries are well balanced, the Eurozone as a whole may still have imbalances.
  • A European supervisor should isolate banks from domestic pressure to pile up sovereign bonds. This can be done by introducing a cap on domestic bonds. Alternatively, diversification of sovereign risk can be forced through the introduction of synthetic Eurobonds or ESBies.

Let me be clear: the basic case for a European banking union is a strong one. A monetary union without a complementing banking union exacerbates systemic risk – as the current situation in the Eurozone painfully illustrates. Monetary unions are in particular not well equipped to deal with asymmetric shocks as regions cannot simply devalue in response to negative shocks. In addition, in a financially well integrated monetary union shocks tend to be exacerbated because agents can easily move capital to other regions, essentially leading to capital flight from the affected regions.1 This worsens the positions of the banks in these regions. To make things worse, the potential of national governments to intervene is limited as the fate of banks and sovereigns is likely to be heavily intertwined in such situations.

A fully-fledged banking union has the ability to address these shortcomings:

  • A European deposit insurance system reduces the risk of capital flight from affected regions and can thus stabilise the Eurozone.
  • Banking resolution at the European level takes away responsibility from national supervisors who might be captured, or simply unable to recapitalise banks due to a lack of resources.
  • Explicit rescue funds for banks break the vicious feedback loop between private (bank) debt and sovereign debt.2
  • A European lender of last resort (if part of the banking union) reduces the risk of self-fulfilling liquidity runs spreading across the Eurozone.

However, a banking union will also create new challenges for systemic risk.

A first problem is that any measure that solely pools national resources at the European level (through a European banking resolution fund or a European deposit insurance for example) can lead to an increase in systemic risk.3 Consider a simple example of two banks, located in country A and B respectively. Suppose that a bank fails if the value of its assets, Ai (i=A,B), falls below its liabilities, D. Suppose that each country also has resources R­i to inject into its bank (either through bailouts or through the national deposit insurance fund). In the absence of a European banking union, the banking system of country A will hence fail if AA + RA < D. Likewise, country B’s bank will fail if AB + RB < D. Eurozone-wide crises will hence occur if AA + RA < D and AB + RB < D. Consider now a full pooling of resources, for example through the creation of a European-wide deposit insurance fund. This will fully eliminate isolated bank failures. Systemic crises will now occur when the joint resources of both countries fall short of the liabilities (when AA + RA + AB + RB < 2D). Such crises necessarily occur more often as now a shortfall at one country can also drag down the bank in the other country.4

What does this imply for the creation of a European banking union? For one, simply merging national deposit insurance systems is unlikely to be an optimal outcome. A preferred system is a two-tier approach with both national and European insurance in place. The national insurance system will be the first line of defence against domestic crises. The European fund (drawing from the national funds of other countries) will only intervene if the national fund is exhausted and when doing so does not undermine its ability to cope with problems in the other countries. Such a conditional insurance system can avoid the negative spillovers associated with a simple pooling – even if no new funds are committed to the system. Note also that this system reduces moral hazard as in the majority of cases the costs will be borne domestically. However, such a scheme cannot effectively address systemic failures as the combined resources are not increased. If politically feasible, a European-fund should thus be equipped with additional resources to deal with systemic events.

A second challenge lies in the harmonisation of supervision and regulation that is likely to come about with a banking union. While such harmonisation is desirable from the viewpoint of eliminating regulatory arbitrage,5 it also poses a great risk. As I have argued earlier (see Goodhart and Wagner 2012), a financial system that is resilient to systemic shocks needs diversity. If all institutions are subject to the same supervisory and regulatory environment, they will tend to undertake similar activities and react in similar ways. This enhances the risk of joint failures.6 There is also no reason to believe that a supranational regulator is necessarily less prone to mistakes. (Just imagine if the excessive credit boom prior to the crisis were not constrained to a few countries in the Eurozone – but had taken place across the Eurozone as a whole!) It is hence of paramount importance that a European supervisor – while creating a level playing field – allows for a diversity of institutional structures and strategies. The supervisor should also ensure that there is competition among different approaches to financial intermediation (in particular, bank-focused financial systems should exist alongside market-based ones).

A third challenge is that of systemic imbalances. While a banking union allows for a more effective resolution of systemic crises, it should also be designed to avoid the build-up of systemic vulnerabilities and hence reduce (as much as possible) large scale crises from the outset. For this regulators need to monitor not only the exposures of individual member countries, but also the combined exposure of the Eurozone. For example, while each individual country may be well-diversified in its exposures, there may still be substantial risk if the majority of countries tend to diversify activities by specialising in the same region. As an example, the European Union was overexposed to the US prior to the crisis (which explains the strong contagion effect in the first phase of the crisis) even though most individual member states were fairly diversified (Schoenmaker and Wagner 2012). A focus on systemic risk also means that regulators can encourage diversity by allowing the banking systems of individual countries to have different exposures – as long as this does not create imbalances at the system level.

A final point is banks’ exposures to sovereign risk. As many have noted, banks had over-accumulated governments bonds of their own countries prior to the crisis. This was a key factor in the amplification of the Eurozone crisis. If, for example, Greek banks had held a well-diversified portfolio of sovereign bonds, the spillover from Greece’s sovereign debt problem to its banking system would have been limited. The problem of imbalanced sovereign exposures intensified during the crisis since banks used LTRO-financing and other rescue measures to increase exposure to their (troubled) sovereigns. This was either because banks underpriced the resulting risk7 or because of pressure from national governments and central banks.

A European regulator can play a key role in limiting this risk factor. He can insulate banks from national sovereign risk by encouraging more diversification of sovereign exposures. This could be done through the introduction of a simple cap on domestic sovereign bonds. A more complete approach would be the introduction of synthetic Eurobonds,8 which force effective diversification of sovereign bond holding in the Eurozone. Such bonds could be promoted by giving them lower risk weights in the calculation of capital requirements or by letting them become the collateral of first choice at the ECB.


In sum, a European banking union has the potential to reduce systemic risk at various margins. However, it also brings about its own challenges. The good news is that those can be largely avoided by a clever design of the institutions that underpin the banking union and by ensuring that regulation and supervision have a truly systemic focus.


Acharya, Viral V (2009), “A theory of systemic risk and design of prudential bank regulation”, Journal of Financial Stability, 5:224-255.
Beck, T, D Gros and D Schoenmaker (2012), “Banking union instead of Eurobonds – disentangling sovereign and banking crises”, VoxEU.org, 24 June.
Beck, T, H Uhlig and W Wagner (2011), “Insulating the financial sector from the European debt crisis: Eurobonds without public guarantees”, VoXEU.org, 17 September.
Brunnermeier, Markus K, Luis Garicano, Philip R Lane, Marco Pagano, Ricardo Reis, Tano Santos, Stijn Van Nieuwerburgh, and Dimitri Vayanos (2011), “European Safe Bonds: ESBies”, Euro-nomics.com.
Dell’Ariccia, Giovanni and Robert Marquez (2006), “Competition Among Regulators and Credit Market Integration”, Journal of Financial Economics, 79: 401-30.
Goodhart, Charles and Wolf Wagner (2012), “Regulators should encourage more diversity in the financial system”, VoxEU.org, 12 April.
Kalemli-Ozcan, Sebnem and Elias Papaioannou (2012), “Banking integration: Friend or foe?”, VoxEU.org, 26 September.
Schoenmaker, Dirk and Wolf Wagner (2012), “Cross-Border Banking in Europe and Financial Stability”, mimeo Duisenberg School of Finance.
Shaffer, Sherrill (1994), “Pooling intensifies joint failure risk”, Research in Financial Services, 6:249-280.
Wagner, Wolf (2010), “Diversification at Financial Institutions and Systemic Crises”, Journal of Financial Intermediation, 19:373-386.

1 While financial integration can help to deal with asymmetric productivity shocks, it may also amplify shocks to the banking system. See Kalemli-Ozcan and Papaioannou (2012).
2 This has been forcefully argued in a recent Vox column (Beck et al. 2012).
3 See Shaffer (1994) or Wagner (2010).
4 With uniformly and independently distributed asset returns Ai, it is easy to see that the likelihood of systemic crises doubles.
5 Harmonisation, in addition, reduces competition among supervisors that can otherwise result in inefficiently lax regulation (Dell'Arricia and Marquez, 2006).
6 Even without homogenous regulation, banks are likely to undertake too similar activities. See, for example, Acharya (2012).
7 Banks do not perceive the full cost of taking on additional sovereign risk since in the case of a failure of their sovereign they may fail anyway.
8 See Beck et al. 2011, and Brunnermeier et al. 2012 for a similar proposal.

1,155 Reads