VoxEU Column Global crisis International Finance

Sovereign default risk and banks in Europe’s monetary union

EZ banks are more exposed to their own nation’s government bonds than ever. This column argues that Eurozone members can now afford to tell their banks to diversify, but pressure from Germany, Austria, France and the ECB might be necessary. Defusing the pernicious entanglement between the Eurozone’s weak banks and weak sovereigns would reduce the cost of any new crisis and reduce the likelihood of such a crisis occurring.

With the one year anniversary of ECB President Draghi’s announcement to “do whatever it takes” and the announcement of the OMT (Outright Monetary Transactions) programme, bond yields have declined and fears of sovereign defaults have receded in countries such as Portugal, Spain and Italy. We all hope that this danger has passed!

But it may well be that we are just witnessing a brief calm before the storm returns. Right now, then, is a good opportunity to solidify what we have and to make the financial system in the Eurozone more robust against such risks in the future.

Diversify bank risk

It always seemed obvious to me that one good route here is to have banks diversify away from national risks. If Spanish banks would not hold Spanish bonds, if Portuguese banks would not hold Portuguese bonds and so forth, then a sovereign default or even just a decline in the prices of these bonds due to heightened default fears would not affect these banks directly.

A similar argument can be made with respect to securities and loan portfolios that may be strongly affected by such default fears. A sovereign default would still lead to losses somewhere, but if these bonds are held as part of a larger portfolio by entities that are capable of bearing some market risk for the reward of the appropriate premium, it seems unlikely that the sovereign default fears will return with the same force, or that a sovereign default will trigger a financial meltdown, as has been feared in the past.

Right now then is a good time indeed to urge the banks to diversify. Or perhaps they have diversified already?

Diversification in practice

Unfortunately, diversification does not appear to have occurred. Research by Acharya and Steffen (2013) document, that “over time, there is an increase in 'home bias' – greater exposure of domestic banks to its sovereign’s bonds”.

Figure 1 provides a simple summary. It is based on data from the European bank stress test in 2011 to show the fraction of sovereign debt held in the form of domestic sovereign debt, aggregating across the banks in the data sample in each country. Banks in Portugal, Spain and Italy hold more than 70% of their sovereign-bond portfolio in terms of domestic bonds. Put differently, rather than safeguarding the future of the European financial system, it has become more fragile. How could this have possibly happened? Why have policymakers not done more to cut this tragic link between home banks and sovereigns, in order to safeguard both?

Figure 1.

A number of reasons may come to mind. In my own paper (2013) I argue that it is the interplay between banks, their regulators and the possibility to shift risks upon the central bank, willing to repo risky sovereign bonds, which is at the heart of the matter. I argue that the sizes of the haircuts applied in the liquidity-provisions by the ECB are irrelevant, given that the ECB is providing the necessary liquidity in the first place.

The paper is stated in abstract terms, but as it is motivated by current events, let me be specific. Imagine that the spectre of sovereign defaults returns. Or, instead, consider the situation a year or two ago, when the crisis was at its peak and defaults seemed within the realm of possibility. If Spain, say, defaults eventually then the banks holding its debt would be particularly affected. With the home bias documented above, these banks would now be Spanish banks.

Conversely, since deposit insurance and bank regulatory mechanisms are national in nature, Spanish bank defaults would adversely affect the finances of the Spanish government.

  • Default risk at Spanish banks in turn increases default fears for the Spanish sovereign.

As governments in high-yield countries became increasingly desperate to find buyers for their debt, their bank regulators encouraged their home country banks to invest in this debt, perhaps in exchange for looking the other way regarding balance-sheet problems or perhaps with other forms of gentle pressure. Bank regulators in safer countries sought to discourage their own banks from holding risky country debt and imposed that any losses had to be borne by bank-equity owners only.

This is a version of financial repression, as described by Carmen Reinhart (2012).

  • As a result – and given the liquidity measures of the ECB – an eventual default in a risky country ends up shifting some of these losses onto the balance sheet of the ECB.

These shifted losses are not priced in, when home country banks purchase home-country debt: consequently, risky countries get to borrow more cheaply, explaining the initial regulatory outcome.

The formal thinking

The formal model is highly stylised in order to focus attention on the core economic linkages: it has three periods; there are many countries, one of which is risky; countries issue sovereign debt in the first period to finance government spending; and there is a common central bank.

  • There are banks in all countries, who have some own equity and who accept deposits in the first period.

They use these funds to purchase sovereign debt from safe as well as risky countries.

  • National regulators in the first period impose bounds on how many risky-country bonds their banks may purchase.

For simplicity, I assume that banks are risk neutral. They choose the optimal portfolio in period 1, given the regulatory bounds as well as the option to default in period 3. Should a bank default in period 3, then its own government will bear a fraction of these losses, thus incentivising the regulators in period 1.

  • In the second period, depositors withdraw their funds.

To satisfy these withdrawals, banks turn to the common central bank and obtain funds through repurchase agreements. The central bank applies a haircut compared to market prices. I assume that the haircut parameter is set such that banks have enough resources to service the second-period withdrawals. This is motivated by the observation that, as of 2013, the ECB has avoided large-scale banking failures through a variety of liquidity-providing policy measures, see Sinn (2012).

  • In the third period, the risky country will receive randomly drawn resources (i.e. an economic shock).

It will use these resources to pay back its debt as well as pay for its share of any national bank failures. If resources remain, it will consume them. I assume that the risky country government and thus, its regulators, seek to maximise this residual consumption. Banks collect debt repayments and repurchase their bonds from the common central bank to the degree possible.

  • If the country defaults and its debt is held by its own banks, and they also default, then the central bank will only partially recover the resources provided in the period-2 repurchase agreements.
What the logic shows

With the theoretical analysis as well as numerical examples, I show two key results:

  • First, there can be equilibriums where countries sell all their debt to home banks, despite adding to the debt burden in default states;.

This partially shifts any such losses to the common central bank.

  • Second, the haircut parameter applied by the ECB in the second period is immaterial for the outcome – it offers no protection against failures and defaults, given that sufficient liquidity is provided by the central bank in the second period.

The reason is simple. What matters are the total resources needed in the second period. With a larger haircut, the banks simply post a larger number of bonds, but when all is said and done, nothing changes regarding the grand total.

Where do we go from here?

Right now, yields are manageable. Countries can afford to tell their banks to diversify and to sell their domestic sovereign bond portfolio on the open market. If my story is right, regulators in Spain, Italy and Portugal won’t do it on their own. Some pressure from Germany, Austria and France would help, as would pressure from the ECB. Given these pressures, there would be no informational stigma attached to these sales either. It will be a good step forward in avoiding a return of the crisis, and to make its consequences less costly, should it return anyhow. The time to act is now.


Acharya, Viral V and Sascha Steffen (2013), “The 'Greatest' Carry Trade Ever? Understanding Eurozone Bank Risks”, CEPR Discussion Paper 9432.

Reinhart, Carmen (2012), “The Return of Financial Repression”, CEPR Discussion Paper 8947, London.

Sinn, Hans-Werner (2012), “Die Target-Falle: Gefahren für unser Geld und unsere Kinder”, Carl Hanser Verlag, Munich, Germany.

Uhlig, Harald, “Sovereign Default Risk and Banks in a Monetary Union”, CEPR Discussion Paper 9606, German Economic Review, forthcoming.

3,150 Reads