All together now? Arguments for a big-bang solution to Eurozone problems

Daniel Gros 05 December 2010



I hope that everything I write in this column turns out to be irrelevant; I very much hope that it will not be necessary to resort to such drastic actions. Economic logic, however, suggests that it might soon represent the least bad solution to a crisis which keeps getting worse. That said…

The horses have left the stable. Europe’s leaders have announced officially that there might be sovereign defaults in the Eurozone. Now they have no good options left. Governments want markets to believe that defaults will happen only after 2013, but what investor is going to wait patiently to be fleeced in a couple of years? The buyers' strike of peripheral Eurozone debt is thus likely to continue, thus raising the cost of the further rescue operations which are clearly on the horizon. The cost of muddling through is increasing by the day.

It would, of course, also be a mistake to let policy be dictated by short-term gyrations in the bond markets. But one recent development has increased the urgency of acting soon. This is the recent announcement of the Eurogroup of November 28th that the loans of the future "European Stability Mechanism" (ESM) would be senior to private creditors.[1]

As I argue at length in a companion column (Gros 2010), this implies that large bailout programs might actually lead to higher risk premiums because large official bailout programmes would imply that any eventual restructuring loses will be shifted to long-term creditors; short-term creditors will have already been paid off in full. 

Moreover, the punitive interest rate (5.8%) imposed on Ireland now by the EFSF implies that a large official loan makes actually default more likely becausethe growth rate one can now expect for the next years (1-2% only) is clearly below the interest rate. When the numerator (debt service) rises faster than the denominator (GDP, i.e. ability to pay), a snowball effect occurs whereby it is ever more difficult for the country to service its debt (which in the case of Ireland would amount to 75% of GNP; see Eichengreen 2010). 

The problem: Vicious circles

This creates the risk of a vicious circle under which a country that has only a manageable problem might be forced into an EFSF (ESM) programme, which would then make debt service more onerous because of the punitive interest rates. This is likely to induce investors to sell the longer-term debt of the country, which would in turn increase the pressure on the country to accept an EFSF programme. The larger the program, the less would be available in the end for bondholders should the programme not work. This is likely to lead to a further increase in the risk premium. The present strategy of "muddling through" on a case by case basis, but insisting that the future mechanism will be senior to private creditors (and that the latter must expect losses), thus carries a strong risk that more and more countries will be forced into a deadly spiral of increasing risk premiums and ever-increasing financing needs.

The solution

The only way out seems to be a big bang; to deal with all the problem cases in one go. The argument against a restructuring of, say, Greek public debt has always been that this would lead to contagion. But contagion is already a fact of life, and it focuses on countries with real problems. Portugal with its combination of high external debt and poor growth prospects looks like Greece. Spain has the “Irish disease”; a real estate bust that leads to huge losses in the banking system. Every country is different, and some countries (Spain, for example) would under normal circumstances not need a bail out. But these are not normal circumstance, and it is not possible to deal with each country in sequence because each bailout lead the markets to expect the next one. Only a big bang can resolve the impasse.

How should this "big bang" look like? A sudden collective default would of course constitute a "mega Lehman" and would have catastrophic consequences. However, it is entirely possible for the countries in question to make investors an exchange offer while continuing to service their payment obligations. There should thus be no technical default, but simply an offer to bondholders to engage in discussions about debt restructuring accompanied by a concrete exchange offer.

Everybody is different

All countries should thus move at the same time, but every country has different problems, and would make a different offer to creditors. Greece and Spain illustrate the two polar cases:

In the Greek case, the problem is clearly the sovereign. Holders of Greek public debt could be offered a par bond (100% of the nominal, but with a low interest rate and a long maturity). This would ensure that banks (and the ECB) would not have to book immediately huge losses on their accounts.

In addition to the par bond, creditors would be offered GDP warrants under which the government of Greece would offer to allocate a certain percentage of any increment in nominal GDP (after the trough expected for 2010/11) to additional payments to foreign creditors, pro rata their present holdings.[2] If Greece were to pay to foreign creditors about 4-5% of any increment in nominal GDP substantial payments could built up over time, with full (even if late) payment possible if Greece returns to a decent growth path. The annex provides some crude model calculations to this effect. For Portugal, a simple rescheduling might be sufficient.

In the Spanish case, the problem stems from the banks. Nobody can know with certainty how large their losses will be in the end. But this uncertainty drags down the entire country. The banks must thus be sacrificed if the sovereign wants to stay afloat. Holders of bonds of the banks most exposed to the real estate bust would thus be offered a debt for equity swap. The Spanish government would then be free of further large contingent liabilities, and should have no problems servicing its present debt of around 60% of GDP.  

The accounting losses for the holders of Spanish bank bonds might again be limited if the bonds are transformed into subordinated debt with the same face value of the bonds. For holders of the bonds which do not mark to market the accounting losses could then be taken over a longer period. Spanish banks would not be forced into fire sales, and patient investors might limit their losses if the Spanish real-estate sector does recover.

The same should have been done in Ireland. But at this point it would require first the (new) Irish government to renege on the guarantee given by the old one. This will lead to legal problems and would formally be equivalent to a default, but it would restore the solvency of the Irish government, so that no haircut would be needed on Irish government debt. The debt-for-equity swap (as with GDP warrants) allows investors to participate in the upside that would materialise if the assets of the Irish banks and Spanish cajas are really worth as much as the banks and their regulators maintain.

Core governments would of course have to stand ready to recapitalise those of their banks with the highest exposure to the peripheral debt to be restructured.

Quick preparation

All this could be prepared during a special weekend meeting of the European Council (followed by a Eurogroup and probably also an EcoFin meeting).

What about the day after? Although this package should restore the solvency of those governments currently under market pressure there might still be initially turmoil in the markets. However, at this point the ECB would be justified in providing abundant liquidity to the interbank market which should then be free of "zombies". Governments and the ECB would thus agree on a division of a labour:

  • The ECB stabilises the banking system, and
  • The EFSF/ESM (the fiscal authorities) take care of the financing needs of governments.

The funding of the EFSF should then be sufficient to cover the (reduced) financing needs of all four GIPS (Greece, Ireland, Portugal and Spain) countries for quite some time.

Patient execution

The big-bang approach is not without risks. It could be prepared in a weekend, but it would require months of patient negotiations to get bondholders to agree.

This is actually very likely to happen because the offer would be close to current market prices and because a large part – maybe even a majority – of the bonds are in the hands of institutions that should respond to political pressures to accept the deal. 

Could a "hold out" by a minority of bondholders who refuse to accept a deal created endless legal problems? There is a solution to this problem suggested by Buchheit and Guti (2010). Greece and other countries could just pass a "mopping up" law which stipulates that any agreement by a super-majority of bondholders (say two-thirds) is binding on the remainder. This would create immediately a statuary "collective-action clause". The absence of "collective-action clauses" thus does not constitute an insurmountable obstacle to reaching an agreement with creditors, as argued recently also in Nouriel Roubini in the Financial Times.


Muddling through is more attractive in the short run, but it does not lead anywhere when doubts about debt sustainability persist and the market has been destabilised by the announcement that the loans of the new permanent crisis mechanism would be senior to private creditors. 

Restructuring will become virtually impossible once the Greek and Irish programs have run their course. At the end of these programmes the major part of the debt of these countries will be owed towards creditors which regard themselves a senior (IMF and ESM), but still impose interest rates far above growth rates. 

At that point the haircut for the remaining private creditors would have to be enormous should the debt sustainability assessment announced by the Eurogroup come to a negative result. Even a low probability of such a result can destabilise markets today making procrastination expensive.


Borensztein, Eduardo and Paolo Mauro (2002). “Reviving the Case for GDP-Indexed Bonds”, IMF Policy Discussion Paper, PDP/02/10, Washington, International Monetary Fund, September.

Buchheit, Lee and Mitu Guti (2010). “How to Restructure Greek Debt”, mimeo.

Eichengreen, Barry (2010). “Ireland’s rescue package: Disaster for Ireland, bad omen for the Eurozone,”, 3 December.

Gros, Daniel (2010). “The seniority conundrum: Bail out countries but bail in private, short-term creditors?”

Miyamajima, Ken (2006) “Hot to Evaluate GD-Linked Warrants: Price and Repayment Capacity”, Working Paper, WP/06/85, Washington, International Monetary Fund, March.

Technical appendix: What value in GDP warrants?

Simple model calculations show that the value of the GDP bonds (or warrants) could be quite substantial. For a much more detailed approach see Miyamajima (2006).

The basic mechanism would be quite simple: each year Greece would dedicate a certain percentage, called τ, of the increment in GDP over the basis year (2010) to foreign debt service (as no interest would be payable on the discount bonds). This percentage might be fixed at 4% (in the case of Iceland the cap was set at 6%). Given that government revenues amount to about 40% of GDP this implies that about 10% of the higher government revenues generated automatically by GDP growth would have to be diverted to foreign creditors.

The annual payments would then be given by τ (Yt - Y0).

If nominal GDP grows with a constant rate, g, (and starting from 2010 GDP level of Y0) the payments (on the GDP bonds only) would evolve over time following: τ (Y0(1+g)t - Y0) = τY0 ( (1+g)t - 1).

Discounting these flows over time at the interest rate i yields the present value of future payments:

Present value of annual payment (until infinity) = Σt { τY0 [ (1+g)t – 1]/(1+i)t}.

This reduces to:   Present value = τY0 (1+i)[ 1/(i-g) – 1/i].

If Greece were to dedicate 4% of the growth in nominal GDP to foreign creditors and the "market" interest rate is 10% the present value (as a ratio of starting GDP) would be given by: 0.04* (1.1)[ 1/(0.1-g) – 10]. 

For a nominal growth rate of 5% (g =0.05) this would amount to 44% of GDP, enough to make up for a haircut of over 30% on the entire stock of Greek debt (now close to 130% of GDP).


[1] Moreover it appears that the loans extended to Greece and Ireland in the context of the existing programs would ex post be transformed into ESM program loans and would thus also become senior.

[2] The case of Iceland provides a recent precedent for GDP linked payments. In this case the governments of England and the Netherlands agreed that Iceland should have paid (at most) 6% of any increment in GDP. However, this deal was not approved in a referendum. Argentina constitutes another precedent, however, in this case this might be rather an example of how to mis-use this concept in the sense that the GDP warrants were not intended to give the investors the potential to recoup their investment fully even in case the Argentine economy recovered fully.

This idea to link payments to capacity to service debt, called Besserungschein in German, is widely used in private transactions. Robert Schiller has called for GDP related financial instruments which would allow borrowers and investors to hedge against shocks to growth. See also Borensztein and Mauro (2002).




Topics:  Global crisis

Tags:  ECB, Eurozone crisis, sovereign restructuring

Director of the Centre for European Policy Studies, Brussels


CEPR Policy Research