This column is a lead commentary in the VoxEU Debate "Euro Area Reform"
In the opening contribution to the VoxEU debate on euro area reform, we and twelve other French and German economists proposed a package of reforms aimed at increasing the stability of the euro area (Bénassy-Quéré et al. 2018). These pursue three objectives:
- The first is to eliminate the ‘doom loop’ between sovereigns and national banking systems, by combining the introduction of a European deposit insurance with concentration charges forcing banks to diversify their sovereign exposure and exploring the creation of a euro area safe asset.
- The second is to improve the resilience of members to adverse shocks, by making European fiscal rules less pro-cyclical and introducing risk sharing mechanisms such as a European unemployment re-insurance.
- And the third is to ensure that if all else fails – fiscal rules and common banking supervision in preventing a crisis, and risk-sharing mechanisms in mitigating its blow – the worst that can happen is an orderly debt restructuring inside the euro area, as opposed to a euro area exit.
As with any reform of the euro financial architecture, the benefits and risks of our proposals are not equally shared across members. Common deposit insurance and fiscal risk sharing could, in principle, come at the expense of countries that are less likely to require fiscal support, such as Germany or the Netherlands. Raising the credibility of debt restructuring could raise the borrowing costs of countries with higher debts and sovereign risk spreads, such as Italy or Greece. We attempt to mitigate these problems in two ways: through the design of individual reform proposals, which seek to minimise these risks and ensure that each country’s core interests are respected; and by ensuring that every single euro area country would be amply better off if our proposals were implemented as a package. We also emphasise that the transition to the new regime should be carefully managed – for example, by exempting bonds currently held in bank portfolios from concentration charges.
As the contributions to the VoxEU debate attest, not everyone agrees that we have succeeded. Peter Bofinger rejects our scheme out of hand. Lars Feld argues that our proposal for fiscal risk sharing does not sufficiently deal with moral hazard concerns. Sebastian Dullien, Guntram Wolff, Gregory Claeys and several others argue that our risk-sharing ideas do not go far enough, particularly in the fiscal area. Jérémie Cohen-Setton, Shahin Vallée and Vesa Vihriälä find that we should have said more about the role of the ECB. Finally, several Italian colleagues, including Lorenzo Bini-Smaghi, Stefano Micossi and Guido Tabellini, consider that our proposals to reform ESM policies and regulate sovereign exposures of banks would create unacceptable stability risks for Italy, and hence for the euro area as a whole.
We are particularly concerned by this last line of criticism, for two reasons.
- First, it gives us a failing grade on an issue that we are sensitive to and aimed to address. We, too, are weary of the potentially unintended consequences on high-debt countries of increased market discipline, and we fully agree that sovereign defaults can have devastating costs and should never be considered a routine instrument for resolving debt crises.
- Second, these critics question less the specifics of our reform ideas, or even the balance of our reform package, but the fundamental desirability of two of our three reform objectives – breaking the doom loop, and making debt restructuring a viable option if all else fails. Eliminating these two prongs would result in the collapse of what we consider a logically balanced approach.
In the remainder of this column, we focus on these two broad issues, reiterate the logic of what we are proposing, and respond to these critics.
Debt restructuring as ‘ultima ratio’
One aim of our proposals is to ensure that unsustainable debts can actually be restructured (or at least rescheduled) as a last-resort option. To achieve this, we propose to lower the output costs and financial stability risks of such restructurings for the countries involved and the euro area as a whole, through stronger safety nets (euro area deposit insurance, and a fiscal rainy-day fund) and by limiting the exposure of banks to any individual sovereign. We also propose changes to bond contracts, and possibly the ESM treaty, to allow debt restructuring to become legally binding with the agreement of a supermajority of creditors and shield the debtor from legal risks. Finally, we advocate creating liquidity lines with the ESM, so that prequalified countries retain access to liquidity even when confronted to adverse market reactions.
Unlike others (see Zettelmeyer 2018 for a survey), we do not suggest ‘hard’ commitment devices, such as an automatic rollover of maturing bonds whenever a country receives conditional assistance, or a debt ratio above which ESM support would require a debt restructuring. Commitment devices of this kind could force countries to restructure unnecessarily and induce financial panics when the criteria that would require a restructuring are close to becoming binding. Instead, we propose an ESM lending policy analogous to the policy currently used by the IMF in deciding whether to extend large-scale loans to its members. If the ESM staff finds, based on a pre-agreed methodology, that a crisis country is unlikely to be able to repay its debt, then a debt restructuring should become a condition for ESM financial support. In doubtful cases, a maturity extension should be sought, maintaining the option of a deeper restructuring in the future.
In his criticism of our proposal, Guido Tabellini (2018) makes five arguments. First, as a matter of logic, he claims that debt restructuring should become more likely in our proposal, raising the costs of borrowing for high-debt countries. Second, the change in ESM policy that we are contemplating could be abused politically; hence, “any proposal to reinforce the role of the ESM should go hand in hand with proposals to transform it from an inter-governmental body into a full-blown European institution that operates under majority rule and is accountable to the European (rather than national) parliament”. Third, a sovereign debt restructuring mechanism along the lines that we propose is unnecessary, because most euro area debt is issued under national law. Fourth, our approach to make debt “more easily defaultable” is wrong because “it amounts to a political decision to break a sovereign promise. As such it has devastating consequences on the legitimacy and functioning of democratic institutions.” Finally, and in part for this reason, solvency problems of sovereigns would be better addressed by adding “explicit verifiable contingencies (such as with indexation clauses to nominal GDP)” to debt contracts.
Contrary to Tabellini’s first assertion, our proposals do not imply (much less require) that debt restructuring will become more likely. Conditional on a situation of unsustainable debt, the probability of a debt restructuring at the expense of private creditors should indeed rise. At the same time, the chances that debt becomes unsustainable would be reduced by the fiscal risk-sharing mechanisms that we propose, the presence of a well-designed euro area deposit insurance, and the fact that the prospect of losing market access – if markets believe that sovereign debts are on an unsustainable trajectory – may trigger an earlier fiscal adjustment, without a debt restructuring, than is presently the case. All that would likely reduce sovereign risk premia.
Furthermore, the possibility of orderly debt restructuring would make it much less likely that a deep debt crisis results in an exit from the euro area – because resolving it inside the euro area would require less austerity, and because it would avoid a situation in which the ECB may be forced to choose between discontinuing support to the banking system of an insolvent country and accepting de facto fiscal dominance. Sovereign borrowing costs of euro area members reflect not only restructuring risks but also redenomination risks. Our proposals reduce the latter.
Hence, the net effect of our proposal is more likely to be a reduction than a rise in borrowing costs, even from the perspective of today’s high-debt countries. To manage any risk to the contrary, a transition period for the proposed changes in ESM policies could be defined, giving high-debt countries time to avoid any jump in borrowing costs by demonstrating that their debt ratios remain on a downward trajectory.
With respect to Tabellini’s third point, he is right that most euro area sovereign debt is presently issued under local law, but wrong to suggest that this is a substitute for a pre-defined orderly procedure to restructure debts that have become unsustainable. True, local law was used to facilitate Greece’s 2012 restructuring (especially through setting a supermajority threshold for decisions that are binding on all debt-holders), but in so doing it set the basis for a European framework that should apply to similar cases if and when they occur. Furthermore, euro area bonds issued after 2012 carry collective action clauses that make it more likely that restructurings based on acts of the local legislature will be subject to legal challenge (Grund 2017). More importantly, using the local legislature to in effect change the terms of bond contracts and depart from commonly agreed principles would constitute precisely the kind of “political decision to break a sovereign promise” that Tabellini rightly seeks to avoid.
The fourth of Tabellini’s points mischaracterises what we are proposing. Far from promoting default, we seek contractual and treaty changes that would make it possible to restructure debt within the law – for example, through the ‘one-limb’ collective action clauses proposed by the International Capital Markets Association, promoted by the IMF and endorsed by France and Germany, but not yet adopted in the euro area. Furthermore, we envisage debt restructuring only in rare circumstances – namely, when governments cannot repay, even assuming ample financing and time (e.g. three years) to make policy reforms. These proposals seek to minimise the inevitable reputational loss to governments and democratic institutions that arises in a deep insolvency.
We are more sympathetic to Tabellini’s remaining two points. The governance of the ESM should indeed be reformed to enable its staff and management to take independent decisions, subject to policies agreed by its board. However, transforming the ESM into “a full-blown European institution that operates under majority rule and is accountable to the European (rather than national) parliament” should not be put as a precondition. The creation of the ESM resulted from a refusal by member states and their national parliaments to assign the corresponding fiscal responsibilities to the EU level. Its transformation is unlikely in the absence of democratically legitimate fiscal decision making at the European level. And if we had that, the approach of our paper – which tries to improve the euro area fiscal and financial architecture in a setting in which national parliaments are not willing to cede much control – may not be needed in the first place. Integrating the ESM into the EU framework also presupposes reforming the European institutions, as the larger member states feel diluted in the current institutional set-up and are unlikely to unilaterally cede unilaterally the power they enjoy in the ESM.
We are also sympathetic to GDP indexation and other forms of explicit contingencies in sovereign bonds. But such bonds are not a panacea. New research shows that they would offer little protection against spikes in the debt ratio unless they constitute a large share of the debt stock – and even then, only in countries in which growth volatility is high relative to the volatility of the primary balance (Acalin 2018). This is not true for most euro area countries. Hence, even in the presence of full GDP indexation of its bond stock, the euro area would require a backstop for dealing with unsustainable debt cases.
Regulating banks’ sovereign exposures
A cornerstone of our proposals is to introduce regulation that penalizes concentrated exposures of banks to specific sovereigns (but not aggregate sovereign exposures, provided they are diversified). The stability rationale is to weaken the ‘doom loop’ – the transmission of sovereign risk to banks, who will pass it on to private economic activity (via reduced lending) and eventually back to the sovereign. One can also justify the need to regulate sovereign exposures through a limited liability argument of the type that Tabellini uses to explain why bank exposures to the domestic sovereign shot up in the crisis – the risks that concentrated sovereign exposures of banks pose to the financial system are not fully internalised by banks purchasing risky debt.
Tabellini argues that regulation of this type would make no difference, because “[a]ny bank is unlikely to survive the default of its sovereign, irrespective of how much domestic debt it holds. A debt default is a dramatic event accompanied by extreme economic and political disruptions. A defaulting country could choose to exit the Eurozone.” We agree that it is possible to imagine a default so cataclysmic that it would create havoc in the entire banking system. But the purpose of our proposals is to avoid such cataclysmic defaults, instead seeking orderly, ‘pre-emptive’ debt restructurings that do not force the country to leave the euro area. Furthermore, Tabellini’s claim is not correct as an empirical matter. To start with, Greece is a counterexample; more broadly, only 17 of 91 sovereign defaults between 1978 and 2005 considered by Kuvshinov and Zimmermann (2017) – defined as failure to make a payment on the due date contained in the original terms of issue – coincided with systemic banking crises. Systemic banking crises following pre-emptive restructurings – where no payments are missed – are even rarer (Asonuma et al. 2018). When such crises did occur, however, they greatly raised the costs of default or restructuring, underlying the need to avoid banking crises as a condition for orderly restructuring.
Citing a thoughtful speech by Vitor Constâncio (2018) and a stimulating study by Bofondi et al. (2018), Tabellini also argues that direct exposure of banks to sovereigns had no significant effects on the transmission of sovereign risk during the European crisis. However, Constâncio (2018) is careful to acknowledge that “[i]ndisputably, the credit risk situation of sovereigns affects banks via several channels, including the amount of debt they hold”. Indeed, Constâncio’s assertion that “CDS premia do not even show that banks with higher ratios of domestic public debt did significantly worse than others with lower ratios” is contradicted by two papers that examine this evidence directly (Beltratti and Stulz 2015, Schnabel and Schüwer 2016). Bofondi et al.’s (2018) finding that direct exposure to the sovereign made no difference in explaining lending behaviour of banks operating in Italy in 2011 is striking, but it appears to be an outlier. Using somewhat different datasets, Gennaioli et al. (2014a,b) and Altavilla et al. (2017) find that higher sovereign exposures of banks did indeed magnify the impact of sovereign distress on the real economy.1 A possible interpretation is that Italy was indeed a special case in 2011 – prior to the OMT announcement – in which debt restructuring and euro area exit were viewed as joined at the hip. But with an appropriate crisis management framework, there is no reason why a debt restructuring, particularly a pre-emptive one, should result in euro exit.
The core of our disagreement is Tabellini’s claim that access to the domestic banking system will remain an indispensable ‘safety valve’ as long as the euro area is not equipped with a lender of last resort for sovereigns. This amounts to saying that, short of a fundamental change to the mandate of the ECB, no collective safety arrangement can substitute the privileged relationship between a sovereign and the national banks. The logical conclusion from this view is that Banking Union is a misguided project. It is also a perfect justification for ring-fencing and avoiding the sharing of risk. Whereas we perfectly understand, and even share the Italian argument that the national ‘safety valve’ cannot be abandoned if not substituted by strong and credible collective safety arrangements, we fail to see how a monetary union where each sovereign’s ultimate creditor remain its banking system would avoid falling again into the same type of crisis it experienced in 2011-2012.
This all said, we agree that direct exposure to sovereigns is clearly not the only channel through which sovereign stress is transmitted to banks. Furthermore, diversification will only partly shield the banking system even from direct exposure (Alogoskoufis and Langfield 2018), which is one of the reasons why in Bénassy-Quéré et al. (2018) we favour the introduction of a euro area safe asset as a complement to regulatory reform. Regulatory penalisation of concentrated sovereign exposures is not sufficient to eliminate the doom loop, but it is certainly necessary.
Guido Tabellini and other analysts are right to insist that any attempt to raise market discipline in the euro area needs to be mindful of its potentially destabilising effects. But they are wrong to conclude that for this reason, it is best to continue with an approach in which the sovereign effectively borrows from its national banking system in a distressed situation, and the handling of crises involving unsustainable debt remains unspecified.
- Failure to outline a procedure for the orderly restructuring of unsustainable sovereign debts implies that in a crisis, these countries would continue to be ‘rescued’ in the standard way, through a conditional loan by the ESM. To avoid openly violating the ESM’s charter, which prohibits such loans, the EU would need to pretend that – with sufficient fiscal adjustment – the debts of the crisis country are in fact sustainable. As witnessed by the failed first programme with Greece, this would lead to exceptionally harsh austerity, high unemployment, and large output declines. In spite of these efforts, restoring solvency will eventually require significant official debt relief.
- Continued reliance on the national banking system as an informal lender of last resort would enshrine the ‘doom loop’ between sovereigns and domestic banks. It is inconsistent with a genuine banking union, in which the ‘nationality’ of banks within the euro area should not matter. Instead, it would be better to strengthen official liquidity provision in a crisis. While the combination of ESM conditional assistance and the OMT goes a long way towards addressing the euro area’s lack of a lender of last resort, it does not help countries that would like access to emergency liquidity without having to negotiate a formal conditional support program. For this reason, we argue for precautionary credit lines that a country such as Italy could qualify for provided that its fiscal and other policies remain on track.
Bailouts of countries with unsustainable debts and reliance on national banking systems as a ‘safety valve’ are inconsistent with a stable, financially integrated euro area. They will weaken the resilience of the euro area and eventually lead to euro exit either of creditor countries that do not want to pay the redistributive price, or of a crisis country that is being subjected to unreasonable austerity. The rise of anti-euro parties in central and northern Europe and Greece’s near-exit in 2015 are warning signals that should not be ignored. It is time for the euro area to address its remaining vulnerabilities head on.
Acalin, J (2018). “Growth-indexed Bonds and Debt Distribution: Theoretical Benefits and Practical Limits”, Working Paper 18-7, Peterson Institute for International Economics.
Alogoskoufis, S and S Langfield (2018), “Regulating the doom loop”, ESRB Working Paper No 74.
Altavilla, C, M Pagano and S Simonelli (2017), “Bank exposures and sovereign stress transmission”, Review of Finance 21(6): 2103-2139.
Asonuma, T, M Chamon, A Erce and A. Sasahara (2018), “Costs of Sovereign Defaults: Restructuring Strategies, Bank Distress and the Credit-Investment Channel”, mimeo, IMF.
Beltratti, A and R M Stulz (2015). “Bank sovereign bond holdings, sovereign shock spillovers, and moral hazard during the European crisis”, NBER Working Paper No. 21150.
Bénassy-Quéré, A, M Brunnermeier, H Enderlein, E Farhi, M Fratzscher, C Fuest, P-O Gourinchas, P Martin, J Pisani-Ferry, H Rey, I Schnabel, N Véron, B Weder di Mauro, and J Zettelmeyer (2018), “Reconciling risk sharing with market discipline: A constructive approach to euro area reform”, CEPR Policy Insight No. 91.
Bofondi, M, L Carpinelli and E Sette (2018) “Credit Supply During a Sovereign Debt Crisis”, Journal of the European Economic Association 16(3): 696-729
Constâncio, V (2018) “Completing the Odyssean Journey of the European Monetary Union”, Remarks at the ECB Colloquium on the Future of Central Banking, Frankfurt, May
Gennaioli, N, A Martin, and S Rossi (2014a), “Sovereign default, domestic banks, and financial institutions”, Journal of Finance 69(2): 819-866.
Gennaioli, N, A Martin, and S Rossi (2014b), “Banks, Government Bonds and Default: What do the Data Say?” CEPR Discussion Paper No. 10044.
Grund, S (2017). “Restructuring government debt under local law: the Greek case and implications for investor protection in Europe”, Capital Markets Law Journal 12(2): 253–273.
Kuvshinov, D, and K Zimmermann (2017), “Sovereign Going Bust: Estimating the Cost of Default,” mimeo, University of Bonn.
Schnabel, I and U Schüwer (2016), “What Drives the Relationship between Bank and Sovereign Credit Risk?”, German Council of Economic Experts Working Paper 07/2016.
Tabellini, G (2018), “Risk Sharing and Market Discipline: What is the Right Mix?”, VoxEU.org, July.
Zettelmeyer, J (2018), “Managing Deep Debt Crises in the Euro Area: Towards a Feasible Regime,” Global Policy 9(Supplement 1): 70-79.
 Based on a panel data set of emerging and advanced countries between 1980 and 2005, Gennaioli et al. (2014a) find that post-default declines in private credit are stronger in countries where banks hold more public debt. Using bank-level data for 20 sovereign default episodes between 1998 and 2012, Gennaioli et al. (2014b) find that there is a large and significant correlation between banks’ bond holdings and subsequent lending reductions. Also using bank-level data, Altavilla et al. (2017) examine the relationship between sovereign stress and bank lending in five euro crisis countries (Greece, Ireland, Italy, Portugal and Spain) and find that “banks’ domestic sovereign exposures in the stressed countries were associated with a statistically significant and economically relevant amplification of sovereign stress transmission to corporate lending, which cannot be attributed to spurious correlation or reserve causality”.