EU banks and sovereign debt exposures

Adrian Blundell-Wignall, Patrick Slovik 14 September 2010



Europe has suffered its own brand of crisis, leaving many of its banks in a mess. One important step identified by economists to cleaning up the banks was to make public stress tests of Eurozone banks (Baldwin et al. 2010). In July, Europe’s policymakers completed and published their stress test for the EU banking system. The results were positive. They showed a sturdy and resilient EU banking system with strong capital positions in the benchmark case and a sound performance in an adverse risk scenario, including an adverse sovereign debt shock. Yet despite these encouraging results, markets have continued to be concerned about sovereign-risk issues and continue to factor in the possibility of default in some countries. Equity markets have also performed poorly and banks remain reluctant to lend. How could this be? The encouraging stress test results seem on the face of it to be at odds with these market developments. This article puts the ongoing market concerns and the EU stress-test results into common perspective.

Sovereign exposures and the stress test

In a recent OECD working paper (Blundell-Wignall and Slovik 2010), we show that while most of the sovereign debt exposures of EU banks are held on banking books, the EU stress tests only considered the smaller trading book exposures. The rationale for excluding the banking books exposures is that over the 2 years considered in the stress test, a sovereign default is virtually impossible given the existence of the European Financial Stability Facility, which is certainly large enough to meet funding needs of the main countries of concern over that period.

The exposures of the EU banking sector to EU sovereign debt are shown in Table 1. The haircuts applied to the trading book in the stress test are shown in the first block of the table. The trading book exposures (not reported in the stress test report) are also shown. The EU-wide loss from the haircut is around €26 billion or 2% of the Tier 1 capital of EU banks at the end of 2009. The contribution of the 5 countries where most of the market focus has been – Greece, Ireland, Italy, Portugal, and Spain – is only €14 billion or 1% of the Tier 1 capital of EU banks.

Yet a different picture emerges when we consider the sovereign debt exposures held on banking book, which are much larger than those of the trading book – around 83% of the total. If the same haircuts are applied to these exposures the loss amounts to €139 billion or 12% of the Tier 1 capital of the EU banks at the end of 2009, and €165 billion or over 14% of Tier 1 capital if trading book losses are included. The haircuts of the 5 countries of market focus amount to €76 billion in the banking book, and €90 billion if the trading book amount is included. This is around 8% of EU Tier 1 capital of stress tested banks.

Table 1. Sovereign debt exposures of EU banking sector: Trading book vs. banking book

Source: OECD, Bank reports.

It is also important to note that the exposures are not evenly distributed – with the banks in some countries more heavily exposed to the countries of focus in the sovereign debt crisis than others. Table 2 shows the exposures of the banks to the sovereign debt of Greece, Hungary, Ireland, Italy, Portugal, and Spain. Both gross exposures and the percentage of Tier 1 capital each country represents are shown (the impact on bank capital can be obtained by applying a haircut assumption to the latter).

The main observations are:

  • Banks tend to be heavily exposed to the sovereign debt of their own country. The exposure of Greek banks to Greek sovereign debt represents 226% of their Tier 1 capital. In Italy, Hungary, Spain, Portugal, and Ireland these numbers are 157%, 133%, 113%, 69% and 26%, respectively.
  • Large cross-border exposures (defined as an exposure above 5% of Tier 1 capital) to Greece are present for Germany, France, Belgium (all with systemically important banks), Cyprus, and Portugal. Large exposures to Portugal are present in Germany and Belgium; to Spain in Germany and Belgium; to Ireland in Germany and Cyprus; and to Italy in Germany, France, Netherlands, Belgium, Luxembourg, Austria and Portugal.
  • Some banking systems are also exposed to non-Eurozone sovereign debt not subject to the European Financial Stability Facility. For example, Austrian, Belgian and German exposures to Hungary are above the notional 5% threshold.

Table 2. National banking exposure to sovereign debt of Greece, Portugal, Spain, Ireland, Italy, and Hungary

Source: OECD and bank reports.

The assumption of no bank failures and no sovereign defaults

The above data raise a number of issues worth discussing in the context of the stress test and market concerns. Equity markets continue to perform poorly, bond spreads remain elevated, and banks remain reluctant to lend. The encouraging stress test results seem on the face of it to be at odds with these developments. This difference is due in part to the assumptions of no bank failures and no sovereign default that are implicit in the stress test.

(a) No bank failures: accounting versus economic interest

If a bank fails, the question of whether sovereign exposures are held in the banking book or in the trading book disappears. If a bank were to fail, the resolution authority would have to realise asset sales in the market to meet demands from depositors and other creditors. The latent losses on the sovereign portfolio in the banking book would be realised. Therefore, shifts in the market values of sovereign debt held on banking books must be relevant for creditors and stakeholders, unless stress-tested banks can be assumed never to fail. Similar comments would apply to other assets, including special-purpose-entity structured products that may currently be worth more in the banking book than in the trading book. Banks have been choosing to hold assets in the banking book and in many cases to reclassify assets to take advantage of better valuations. Shareholders and/or creditors have to make their own assessments about the probability of solvency events in individual banks which could affect their recovery of funds in the event of insolvencies.

(b) No sovereign default

The assumption of no sovereign default over the period of 2010 and 2011 appears to be very reasonable. The European Financial Stability Facility is made up of a €720 billion lending facility (€220 billion from the IMF; €60 billion from the EU; and the special-purpose vehicle can build exposures for 3 years to the limit of €440 billion for the 16 Eurozone countries) which provides a guarantee of funding for any countries facing financing pressures, certainly for the next 3 years.

Table 3 shows the approximate funding needs of Greece, Ireland, Italy, Portugal, and Spain based on simulations using OECD growth and deficit forecasts. The European Financial Stability Facility could more than cover all of the funding needs of these countries, even in the unlikely case that no securities could be sold on the open market, for the period of the stress test (2009-2011). So the assumption of no default over this short period is reasonable.

But what happens after 2012? The concerns in the market beyond this point are a combination of the longer-run fiscal sustainability problem and the difficulty these countries face in achieving the structural adjustments in labour and pension markets necessary to ensure sustainable growth in a period of budget restraint. The fear is that this will not be resolved by the time the support packages run out, and hence portfolio managers may not put the probability of restructuring as zero.

Table 3. Budget financing needs vs. European Financial Stability Facility €440 billion (€720 billion including funding from the IMF & EU)

Source: OECD.


Most of the sovereign debt held by EU banks is on their banking books, whereas the EU stress test considered only their smaller trading book exposures. Market participants, however, do not have the luxury of ignoring the sovereign debt held on the banking books.

  • First, individual bank failures would see latent losses on the banking book realised, a fact that creditors and equity investors need to take into account.
  • Second, and more importantly, the market is not prepared to give a zero probability to debt restructurings beyond the period of the stress test and/or the period after which the role of the European Financial Stability Facility comes to an end.

The main reasons for this fear are clear. Markets are concerned about the mammoth task ahead for fiscal consolidation in a period of weak economic growth, and the apparent difficulty of achieving structural/competitiveness reforms in some countries in a relatively short period of time.


Baldwin, Richard, Daniel Gros, and Luc Laeven (2010) “Completing the Eurozone rescue: What more needs to be done?”, A Publication, 17 June.

Blundell-Wignall, A and P Slovik (2010), “The EU Stress Test and Sovereign Debt Exposures”, OECD Working Paper on Finance, Insurance and Private Pensions 4, OECD Financial Affairs Division.



Topics:  Europe's nations and regions International finance

Tags:  financial regulation, stress tests, Fiscal crisis, Eurozone crisis

Special Advisor to the Secretary-General on Financial Markets; and Deputy Director, Directorate for Financial and Enterprise Affairs, OECD

Economist and Policy Analyst at the OECD


CEPR Policy Research