Disclaimer: The views expressed in this paper are solely those of the author and do not represent the views of the Central Bank of Ireland or of the Eurosystem.
Shortly after the default of Lehman Brothers in 2008 several Irish banks faced acute funding problems. In order to avoid a systemic crisis the Irish government extended a wide-ranging guarantee to its banks. As credit losses mounted in tandem with deteriorating economic conditions, the solvency of the Irish State itself became threatened and eventually compelled Ireland to enter an EU-IMF Programme of Assistance. This paper discusses the alternatives available to European policymakers regarding the regulatory treatment of banks’ sovereign exposures in light of the Irish crisis. Of the various reform options, strict exposure limits seems to be the only regulation that would have materially affected Ireland’s crisis management. However, it is doubtful whether consensus on reform can be achieved.
1. Ireland’s debt crisis
Few countries have witnessed a sovereign debt crisis emerging so quickly and unexpectedly as Ireland did during the great financial crisis. Shortly after the default of Lehman Brothers in 2008 several of the Irish banks experienced acute funding difficulties. Facing a systemic crisis, and convinced that the banking system was broadly solvent, the Irish government issued an extensive guarantee, covering senior debt and some of the existing subordinated debt and covered bonds. Two years later the mounting credit losses threatened the solvency of the Irish State and compelled Ireland to seek an official Programme Assistance from EU and IMF.
It is worthwhile to quickly recapitulate the scale of the Irish property and credit bubble in the decade preceding the crisis – the Celtic Tiger years. Between 1997 and 2007 domestic lenders’ balance sheets grew fourfold, an expansion largely funded by foreign banks and investors. The loan-to-deposit ratio in domestic banks rose from 130 per cent to 220 per cent. Virtually all of the external funds went into the property sector, both commercial and residential. The annual construction of homes grew from a sustainable rate of about 30 thousand units in 1996 to more than 90 thousand ten years later. The stock of mortgage loans exploded from €16 billion in the first quarter of 2003 to a peak of €106 billion by the third quarter of 2008 and household indebtedness peaked at 120% of GDP in 2010.
Meanwhile the State’s finances flourished. Despite public wages were inflating, full employment and tax revenues from the construction sector ensured recurrent budget surpluses. Ireland went into to the crisis with a gross debt-to-GDP ratio of just 24 per cent, one of the lowest in the European Union. Hence, it seemed that Ireland’s public finances in 2007 could withstand any imaginable downturn in the economy.
Things reversed in a dramatic fashion as funding markets first trembled in 2007 and then imploded after the default of Lehman Brothers. Over the next four years, retail property prices fell by half in Ireland, and commercial real estate prices more than so. The blanket guarantee issued by the Irish government in September 2008 had shifted a stock of about 400 billion or 200 per cent of GDP of contingent bank liabilities to the State (including deposits already covered by the national deposit insurance). The guarantee temporarily restored confidence and reversed deposit outflows (partly at the expense of banks in other countries, in particular the UK) but by mid-2010 some 100 Billion euro in retail and wholesale funding had been withdrawn from the Irish banks, which was substituted by central bank lending.
At this point the revised loan loss estimates and mounting budget deficit had started to raise concerns over the State’s ability to service its debts. Following the Deauville statement – in which Angela Merkel and Nicolas Sarkozy announced that private investors should be prepared to take losses in future bank bailouts – the Irish 10-year yield surpassed 8 per cent in November 2010 and Ireland was compelled to apply for an EU-IMF Programme of Assistance. A final assessment of the banks’ credit losses was performed in 2011, which brought the State’s total bank recapitalisation to 64 billion euro.1)For extensive reviews of the Irish financial crisis, see Honohan (2010) and Nyberg (2011).
Ireland’s experience contrasts with that of other peripheral European countries during the great recession, and Greece in particular, where the bank-sovereign loop at least initially worked in the other direction – apprehensions over countries’ fiscal position triggered outflows from their domestic (and sometimes foreign) banks. Although it is clear by now that Ireland’s fiscal model during the Celtic Tiger had become unsustainable, it was undoubtedly the bank rescue that triggered the acute phase of the crisis and the need to seek external assistance.2) The indirect costs of the crisis are likely to be much larger still, probably more than 100 billion. See Honohan (2015). In stark contrast to other stressed Eurozone countries, the amount of domestic government paper held by Irish banks before the crisis was negligible, but rose quickly after 2009. For example, in Italy, Greece and Spain the share of domestic debt of banks’ securities portfolios was above 90% already before the onset of the crisis.3)For analyses of peripheral banks’ holdings of sovereign debt, see, e.g.,Uhlig (2013), Acharya and Steffen (2015) and Altavillo, Pagano, Simonelli (2016) and Affinito, Albareto, and Santioni, (2016). As in other peripheral countries, behind this on-loading of sovereign debt in Ireland, and off-loading of other securities, may be present both aspects of moral suasion and carry trade (as sovereign bonds were high-yielding, risk-weighted at zero and eligible as ECB collateral). However, the predominant reason is much less subtle: Irish government paper was the main instrument used to recapitalise its most loss-ridden lender, Anglo Irish Bank.
Anglo Irish Bank had been growing at a dramatic pace over the past years and had by the onset of the crisis become the third largest lender in Ireland. It was arguably of systemic importance at this point, both in its own right and through potential contagion effects, and was like other domestic banks covered by the blanket guarantee. It was the first bank to recognize major credit losses and was nationalised in January 2009, amidst a write-down of equity and subordinated instruments and a capital injection by the State of four billion euro. In December 2009 the new management of the bank signalled that losses had continued to increase and that they were likely to breach minimum capital requirements. The pressed Irish government, which was fearful of the effects of issuing more debt into the market, issued a Promissory Note – an IOU – with a notional value of 8.3 billion directly to the bank. The value of the Promissory Note was subsequently increased as loss estimates kept rising. In 2010 Anglo was merged with the smaller but equally defunct lender Irish Nationwide, which formed the Irish Banking Resolution Company (IBRC). After a final capital assessment in 2011 the final Promissory Note proffered to IBRC amounted to almost 30 billion euro, more than 15 per cent of Ireland’s GDP, of which almost 26 billion covered incurred losses and loss provisions.4)The around four billion left representing Tier 1 capital made IBRC well capitalized, with a core Tier 1 ratio of 15.1 per cent per December 2011. See the Irish Bank Resolution Corporation (2011).5)In February 2013 IBRC was liquidated and the Promissory Note was exchanged for long-dated (marketable) bonds, to be amortized by the Irish government at a (minimum) pace agreed with the ECB.About one billion of the Note went to the recapitalisation of EBS as it was merged with Allied Irish Bank. The remaining 34 billion of the bank recapitalisation were done in cash, of which two thirds were taken from the National Pension Reserve Fund. See McArdle (2013).
A by-product of the recapitalization was that, since IBRC retained its banking license, the Promissory Note could be used as collateral for emergency lending from the central bank. However, this still proved insufficient to restore confidence in the State’s finances and break the spiral of increasing bonds yields / default risk.
2. Limited Prospects of Regulatory Reform
Much has been achieved to improve the fiscal and financial framework in Europe since the onset of the great recession: The fiscal compact – to strengthen budgetary discipline; the European Stability Mechanism – to act as a backstop lender; the Single Supervisory Mechanism – to strengthen and harmonize supervision of large banks; and the Bank Recovery and Resolution Directive and Single Resolution Mechanism – to enable orderly resolution of banks.
All this should work to reduce the probability of sovereign stress in the long term. But ultimately no degree of fiscal integration or financial sector reform can ever guarantee that sovereign defaults do not occur, only access to (unlimited) monetary financing could do this. When announced in 2012 the ECB’s Outright Monetary Transactions Programme (OMT) was perceived to provide a sufficient backstop.6)Executive Board member of the ECB, Benoît Cœuré (2013) has described OMT as ”an insurance device against redenomination risk, in the sense of reducing the probability attached to worst-case scenarios”. Despite that the programme would only make purchases of sovereign bonds in secondary markets under certain conditions, it was followed by a rapid decline in peripheral countries’ bond spreads and ensured that no more countries had to seek official programmes of assistance alongside Greece, Ireland and Portugal. Although the OMT has not been activated various other lending and purchase programs undertaken by the ECB since 2008 have led to a substantial portion of Eurozone sovereign debt now being held by ECB. Any desire to reduce the ECBs and the European banking sector’s exposure to sovereign debt should arguably start with the recognition that this debt is not risk-free.
One may recall that the preferential treatment of sovereign exposures have been a feature of capital regulation ever since the first Basel Accord in 1988, which provided for zero risk weight for sovereign debt of OECD member countries. Basel 2 and 3 do not include similar provisions but allow, at national discretion, for a lower risk weight to be applied to sovereign (or central bank) exposures that are denominated in domestic currency and funded in that currency. In addition however, the EU framework diverges from the Basel rules as it allows banks to assign zero risk weight to sovereign exposures denominated and funded in the currencies of any member state, while it also exempts EU sovereign debt from limits on large exposures.7)Specifically, any euro area bank may apply a zero risk weight to any euro area sovereign. In addition, there is a transitional arrangement (Article 114(5 that allows a zero risk weight for any bonds issued by any EU member state, irrespective of the currency, until 31 December 2017.)) Due to these deviations the Basel Committee on Banking Supervision (BCBS) has graded the European framework materially non-compliant with Basel III.
While official lenders were spared from the approximately 53 per cent write-down of Greek debt in 2012, virtually all outstanding debt is now held between Greek banks and official lenders. Regardless of whether another write-down of Greek bonds takes place and affects official lenders (as opposed to “merely” reductions of the bonds’ real value), most observers would agree that the regulatory framework as regards banks’ sovereign exposures needs to be changed.
Members of ECBs governing council expressed their frustration already in 2013, when Yves Mersch called the current practice an “illusion that should be stopped”.8)“ECB policymakers say no going it alone for government bond risk weightings”, Reuters, November 26, 2013. However, they also insisted that Europe should not go alone – any reform should be undertaken on a global basis by the BCBS. In March 2015 the European Systemic Risk Board (ESRB) published a report that called for a revision of the regulatory framework without advocating any specific alternative. In November 2015, a Five Presidents’ Report was devoted to severing the link between banks and sovereigns. The report concluded that “zero-risk weighting of sovereign debt in the EU, as well as the exemption from existing large exposure requirements, are a source of vulnerability” (EPSC 2015). The Dutch Presidency has set up an Ad Hoc Working Party to review different options in 2016.
However, it is evident that a proposal to reform the regulatory treatment of (held-to-maturity) sovereign instruments will meet resistance. Let us quickly review the main options for reform.
(i) Risk weights for European sovereign debt are determined in the same way as for other sovereigns, in accordance with the Basel framework.
Removing the carve-outs for European debt in the Capital Requirements Regulation would mean that, in the standardised approach, risk weights would be determined by external ratings – just as they are for sovereign debt of third-party countries and corporate bonds. However, such a proposal would likely face several objections. First, although the impact on the vast majority of banks would be insignificant, some banks would likely be compelled to raise capital.9)For example, an external rating of BBB- implies a risk weight of 50%. Given, say, a target Tier 1 ratio of 10 per cent, it would add just 5 per cent to capital requirement of the nominal amount of sovereign debt held. See Lanotte et al (2016) for an impact assessment on European banks. More importantly, unless Europe could convince the BCBS to implement the reform on a global basis it would mean that European Union (and/or Eurozone) states would be disadvantaged relative to other jurisdictions. Furthermore, since the onset of the financial crisis, and the default of highly-rated mortgage derivatives in the US, it has been part of the policy agenda to decrease reliance on external ratings. Finally, research shows that market signals such as ratings are often imperfect indicators of actual default and tend to be procyclical – particularly so when it comes to sovereign debt (Gros and Mayer 2010).
(ii) Non-zero, administratively set risk weights introduced on either a European or global basis.
This would constitute a compromise allowing policymakers to avoid relying on ratings and to have control over the underlying factors of risk weights – which could include a fixed risk weight for all EU or Eurozone countries. In a way, the leverage ratio introduced by Basel 3 already fulfils such a function. For example, for a bank with a target Tier 1 ratio of ten per cent a leverage ratio of 3 per cent implies a universal risk weight of 30%.
However, it is hard to see that such an approach would gain support. How and by whom would the sovereign ratings be set? Gros (2013) suggests that risk weights be linked to the Maastricht criteria. The risk weight would remain at zero as long as both debt and budget deficit as a percentage of GDP remained below 60 and 3 per cent, respectively, but would be increased if either the deficit or the debt ratio exceeds these values. This seems intuitive but the question is whether this measure would really capture default risk; for example, the proposed method implies that German Bunds would carry a higher risk weight than Lithuanian or Romanian debt. Second, it is hard to see what range of risk weights would be politically acceptable yet economically relevant. In the Basel 2 standardised approach, the lowest non-zero risk weight is 20% (for AAA-rated banks and corporates). Member states may insist that sovereign debt does not attain a higher risk weight than this. Even for the banks with largest holdings of sovereign debt the impact of a risk weight in the order of 10-20% would likely be quite manageable. However, for the very same reason, it would not materially alter the yield/capital trade-off and banks’ incentives.
(iii) Strengthening of Pillar 2 and/or Pillar 3 processes, including stress tests procedures.
In 2011 the European banking Authority took the unprecedented step of guiding banks to hold additional capital for their hold-to-maturity sovereign debt holdings. However, I think similar procedures or a strengthened Pillar 2-treatment are unlikely to be viewed as a sufficient response on their own. It would certainly avoid a difficult political decision, but by delegating the risk assessment to supervisors it would create ambiguity and potentially subsequent disputes. Even if the ECB adopted a politically agreed, hard-wired approach to the banks under its supervision, non-eurozone banks are supervised by national regulators that may have quite differing views as to their own and other sovereigns’ default risk – and how to mitigate it. As regards Pillar 3 measures, banks’ holdings of sovereign bonds are by this stage fairly well known and this transparency (or lack thereof) is not likely to affect behaviours until sovereign default becomes a distinct possibility, as was the case in Greece.10)As is well documented, banks in peripheral countries increased their stock of own sovereign bonds during the financial crisis, not the other way around See Altavilla, C., Pagano, M., and Simonell, S. (2016). The haircuts applied by the ECB, which range from 6 to 16 per cent for BBB-rated sovereign paper (as a function of residual maturity and coupons) are arguably too small to be a binding constraint on banks funding.
(iv) Remove or revise the exemption for sovereign exposures from large exposure rules
An exposure limit would have several advantages. First, it would not put European banks at a capital disadvantage. It would also avoid an excruciating political process of reaching agreement on alternative risk ratings and risk weight mappings. Finally, it is a measure that would be relatively easy to phase in over an extended period of time. The Basel regulations limit single counterparty exposures to 25 per cent of a bank’s own funds. For sovereigns Europe could decide on a higher limit, such as 50 or 100 per cent. One should also recall that the limit would apply to individual sovereigns, not to the aggregate level of a bank’s sovereign debt holdings. Hence, banks that prefer to hold more sovereign debt (e.g. to meet liquidity buffer requirements) could do so by holding other countries’ Diversification of banks’ sovereign debt portfolios could also be explicitly encouraged through the regulatory framework.11) ESRB (2015) suggests that a possible definition of a well-diversified portfolio relates to the corresponding country’s GDP in the euro area GDP. The drawback of exposure rules is of course that they do not differentiate with respect to (perceived) risk, but one can imagine a framework that combines exposure limits with credit risk ratings, so that the limits become stricter or kick in at lower amounts the riskier is the debt in question.
Exposure limits is also the recommendation of the Five Presidents’ Report (EPSC 2015). The Report concludes that “[a] straightforward exposure regime would greatly limit systemic risk in the banking system, result in a well-diversified government debt portfolio and considerably weaken the doom loop between sovereigns and their banking systems”. However, there is already resistance to this idea. Lanotte et al (2016) argue that a 100 per cent concentration limit would force Italian banks to shed about half of their holdings of domestic bonds, which would amount to around 200 billion euro, or 13 per cent of GDP. The authors argue that a long phase-in period of exposure limits may not mitigate the sell pressure on Italian bonds as “markets have shown a strong tendency to front-load any regulatory changes”. They also argue that risks for financial stability are still material and any new regulation should be held off until “normal times”. The ECB’s Vice President Vitor Constâncio has also expressed his disagreement with the idea of introducing exposure limits.12)Constancio (2015).
3. The European Regulatory Framework in Light of the Irish Crisis
The Irish experience illustrates the inherent instability of the current European framework and the single currency. The prohibition of monetary financing and lack of risk-sharing create very strong incentives for fiscal discipline among member states, which leads to a world with distinct equilibria: In the good equilibrium, which persisted until the default of Lehman Brothers, default risk was perceived as low by market participants even though many countries’ economic fundamentals were deteriorating. When the hurdle was passed however, risk premia shot up and redenomination or default, which previously was unthinkable, quickly became a self-fulfilling prophecy.
When the failure of the Irish banks threatened both Irish and European financial stability in 2008, and lacking a mechanism to restructure systemic banks, the Irish government chose to step in as guarantor for its banking system. As the fiscal situation deteriorated in tandem with property prices and credit losses the solvency of Ireland itself became threatened. Since the State could no longer viably borrow in the market to meet the increasing capital needs of its banks it had to resort to recapitalization via direct injection of non-marketable government paper, ultimately amounting to almost 30 billion euro. These IOUs were in turn used as collateral for liquidity support from the central bank. Arguably, the only alternative for Ireland to break the spiral of higher bond yields and higher default risk at that point would have involved a European risk-sharing mechanism, palpably lacking at the time.
How would the management of the Irish crisis been affected by a different regulatory treatment of banks’ holdings of sovereign debt? Before discussing this one may first note that the BRRD, the newly introduced bank resolution regime, would probably not, in itself, have changed the course of events as regards Anglo Irish Bank. For example, if the bank had been required to hold bail-inable debt and equity (MREL) equal to twice its capital requirement it would have increased the bank’s loss absorption capacity by about four billion. Certainly helpful but small relative to its ultimate capital need.13)Based on Anglo Irish Bank’s accounts as of September 30, 2008.
It is also clear that none of discussed reform options other than exposure limits would have stymied the government’s resolution strategy. Suppose for example that Irish sovereign debt had carried a risk weight of 20 per cent (Ireland had an AA-rating in 2009). The same Tier-1 capital ratio of 15 per cent in IBRC that was achieved by a Promissory Note worth 30 billion would then have been achieved with one worth 31 billion instead.14)With a target Tier-1 ratio of 15% and a risk weight of 20%, the 31 billion would have necessitated 960 million in additional equity. On the other hand, even a fairly generous exposure limit would have prevented the recapitalization of IBRC: the resulting ratio of Irish sovereign debt to the bank’s own funds became about 750 per cent. If a strict exposure limit had been in place it is anyone’s guess how Irish and European authorities would have acted back in 2009. In order to avoid a bank failure at the time it is possible that Anglo Irish Bank would have been merged into one of the larger banks, which eventually would have brought about an even bigger problem. It is also possible that some of the European mechanisms that were introduced later on, notably the EFSF, would have been put in place so as to avoid contagion from the Irish crisis. If not, it is likely that Ireland would have been compelled to apply for a Programme of Assistance already in 2009.
However, it is plausible that the recognition of sovereign risk would have induced both national and European policymakers to engage in necessary reforms much earlier. If the function of banks as “sovereign debt buyers of last resort” had been restricted at the outset, it is possible that insolvency procedures for both banks and sovereigns had been developed at the very inception of the Eurozone, thus lessening the likelihood of a debt crisis in the first place.
Acharya, V. and Steffen S. (2015). The Greatest Carry Trade Ever? Understanding Eurozone Bank Risks. Journal of Financial Economics, 115, 215-236.
Affinito, M., Albareto, G. and Santioni, R. (2016). Purchases of sovereign debt securities by Italian banks during the crisis: the role of balance sheet conditions. Banca d’Italia Occassional Paper 330.
Altavilla, C., Pagano, M., and Simonell, S. (2016). Bank exposures and sovereign stress transmission. ESRB Working Paper Series issue 11.
Bank for International Settlements (2014). Regulatory Consistency Assessment Programme (RCAP) – Assessment of Basel III regulations – European Union.
Capital Requirements Regulation (EU) No 575/2013 of the European Parliament and of the Council of 26 June 2013 and Capital Requirements Directive 2013/36/EU.
Cœuré, B. (2013). Outright Monetary Transactions, one year on. Speech, September 2.
Constâncio, V. (2015). The Role of Stress Testing in Supervision and Macroprudential Policy. Speech, October 29.
European Political Strategy Centre (2015). Further Risk Reduction in the Banking Union. Five Presidents’ Report Series, Issue 3.
European Systemic Risk Board (2015). ESRB Report on the Regulatory Treatment of Sovereign Exposures.
Gros,D. and Mayer T. (2010). Create the European Monetary Fund Now. CEPS Policy Brief 202.
Gros, D. (2013). Banking Union with a Sovereign Virus: The self-serving regulatory treatment of sovereign debt in the euro area. CEPS Policy Brief 289.
Honohan, P. (2010). The Irish Banking Crisis – Regulatory and Financial Stability Policy 2003-2008.
Honohan, P. (2015). Where did all the money channeled into property-backed lending go? Irish Times, January 31.
Irish Bank Resolution Corporation (2011). Limited Annual Report & Accounts.
Lanotte, M., Manzelli, G., Rinaldi, A.M., Taboga M. and Tommasino, P. (2016). Easier said than done? Reforming the prudential treatment of banks’ sovereign exposures. Banca d’Italia Occassional Paper 326.
De Marco, F. and Macchiavelli, M. (2014). The Political Origin of Home Bias: the Case of Europe. Unpublished manuscript.
McArdle, P. (2013). Refinancing the Irish Bail-Out – The Options post the June 2012 Summit. IIEA Working Paper Series issue 12.
Nyberg P. (2011). Misjudging risk: Causes of the systemic banking crisis in Ireland. Report of the Commission of Investigation into the banking sector in Ireland.
Uhlig, H. (2013). Sovereign Default Risk and Banks in a Monetary Union. German Economic Review, 15(1), 23—41.
Footnotes [ + ]
|1.||↑||For extensive reviews of the Irish financial crisis, see Honohan (2010) and Nyberg (2011).|
|2.||↑||The indirect costs of the crisis are likely to be much larger still, probably more than 100 billion. See Honohan (2015).|
|3.||↑||For analyses of peripheral banks’ holdings of sovereign debt, see, e.g.,Uhlig (2013), Acharya and Steffen (2015) and Altavillo, Pagano, Simonelli (2016) and Affinito, Albareto, and Santioni, (2016).|
|4.||↑||The around four billion left representing Tier 1 capital made IBRC well capitalized, with a core Tier 1 ratio of 15.1 per cent per December 2011. See the Irish Bank Resolution Corporation (2011).|
|5.||↑||In February 2013 IBRC was liquidated and the Promissory Note was exchanged for long-dated (marketable) bonds, to be amortized by the Irish government at a (minimum) pace agreed with the ECB.About one billion of the Note went to the recapitalisation of EBS as it was merged with Allied Irish Bank. The remaining 34 billion of the bank recapitalisation were done in cash, of which two thirds were taken from the National Pension Reserve Fund. See McArdle (2013).|
|6.||↑||Executive Board member of the ECB, Benoît Cœuré (2013) has described OMT as ”an insurance device against redenomination risk, in the sense of reducing the probability attached to worst-case scenarios”.|
|7.||↑||Specifically, any euro area bank may apply a zero risk weight to any euro area sovereign. In addition, there is a transitional arrangement (Article 114(5|
|8.||↑||“ECB policymakers say no going it alone for government bond risk weightings”, Reuters, November 26, 2013.|
|9.||↑||For example, an external rating of BBB- implies a risk weight of 50%. Given, say, a target Tier 1 ratio of 10 per cent, it would add just 5 per cent to capital requirement of the nominal amount of sovereign debt held. See Lanotte et al (2016) for an impact assessment on European banks.|
|10.||↑||As is well documented, banks in peripheral countries increased their stock of own sovereign bonds during the financial crisis, not the other way around See Altavilla, C., Pagano, M., and Simonell, S. (2016).|
|11.||↑||ESRB (2015) suggests that a possible definition of a well-diversified portfolio relates to the corresponding country’s GDP in the euro area GDP.|
|13.||↑||Based on Anglo Irish Bank’s accounts as of September 30, 2008.|
|14.||↑||With a target Tier-1 ratio of 15% and a risk weight of 20%, the 31 billion would have necessitated 960 million in additional equity.|