Curing Demotion Sickness

By Arnaud Marès | London

Hamiltonian moment. As the market has made abundantly clear, the current stage of the global sovereign debt crisis is not and never was an 'Irish' crisis. It is primarily the consequence of a demotion of government bonds in the liability structure of governments, transforming them de facto, if not de jure, from senior to subordinated debt. This makes it, in our view, more likely that an increasing number of countries eventually have to rely on inter-government funding assistance. Paradoxically, though, splitting government liabilities between senior and subordinated debt can lead to a credible - and lasting - solution to the Euroland sovereign debt crisis, which we outline below. The announcements on November 28, 2010 ought therefore to result in an acceleration in the pace at which the crisis is unfolding and make its outcome more binary: if wide-scale debt repudiation (a debt 'jubilee') is to be avoided, then Europe must likely move faster from the current fiscal confederation towards a fiscal federation. Not for the first time, Europe stands at a crossroads. The test it faces is reminiscent of that which confronted the US at the time of Alexander Hamilton's assumption plan of 1790.

Seniority matters. We raised some time ago the question of where bondholders rank, in terms of seniority, in the liability structure of governments (see "Ask Not Whether Governments Will Default but How", Sovereign Subjects, August 25, 2010). This is a problem that has two layers:

Creditors versus fiscal stakeholders. The first layer is the question of where financial creditors as a group rank relative to fiscal stakeholders, by which we refer to taxpayers and beneficiaries of public spending. Starting from a position where almost no advanced economy's government is on a stable debt trajectory, creditors are senior enough to be sheltered from losses if fiscal stakeholders are able and willing to bear the fiscal pain required to stabilise debt metrics. We discussed this question in a previous note (see "Sense and Sustainability", Sovereign Subjects, November 8, 2010) and underlined the crucial role that interest rates play in determining how high the hurdle of adjustment is, and therefore how economically and politically feasible it is.

Why the interest rate on loans to Ireland is too high. Seen in this context, the disappointment of the market (and of Ireland's opposition benches) at the cost to the Irish government of EU and IMF assistance is not surprising. The question is not whether a rate of 5.82% (the indicative rate assuming that loans are entirely drawn down) is lower than what the government would have to borrow at in the market currently. The question is whether the rate of 5.82% is low enough to make the fiscal adjustment credible, and therefore ensure that debt will be repaid over time.

In itself, a cost of funding of 5.82% is not unbearable for the Irish economy, but it is in our view unnecessarily high. Elga Bartsch, our Chief European Economist, considers that the government's medium-term growth assumption of 2.75% per annum on average is optimistic, but not wildly so (see Euroland Economics: Bailing Out Ireland and Beyond, November 29, 2010). A slightly more conservative assumption for growth, and a conservative assumption for inflation, (implying a 3.5% nominal growth rate in the medium term) would result in the Irish government having to generate a permanent primary surplus of around 3% of GDP over time to stabilise its debt ratio near the projected peak of 125% of GDP. This is taxing (quite literally) but not unachievable by any means.

Why then do we consider this rate to be unnecessarily high? If the purpose of inter-government assistance is to avoid default, then the lower the rate on official loans, the higher the likelihood that the objective is reached. Charging to the borrower a margin on top of the cost of funding of the lenders does not maximise the likelihood that default is avoided. In fact, it does not improve it meaningfully relative to the situation that prevailed up to end-October, when the 10-year yield on Irish government debt approximated 6-6.5%.

Government bonds are no longer the most senior claim on government. Comparing Ireland's cost of funding through peer assistance to current bond yields (8-9% at ten years) is not relevant, because in the meantime the nature of Irish government bonds - and indeed that of all European sovereign bonds - has changed. From senior debt, they have been demoted, de facto if not de jure, to subordinated debt - a key development in the sovereign debt crisis, which we discuss now.

Bond holders versus official lenders. We move here to the second layer of our analysis of seniority and subordination of government liabilities: if a government has to impose outright losses on its creditors, how would that loss be allocated between government bond holders and other creditors? This is the question that has weighed on peripheral European sovereign bond markets ever since the European Council of October 28-29, 2010, when the prospect of imposing haircuts as a condition for inter-government support was first raised.

Eurogroup clarification: welcome but not good news. Since then, the relative ranking of government bonds and official loans in the liability structure of governments has been clarified, in the form of a statement by finance ministers of the Eurogroup on November 28, 2010. While this clarification was welcome, we believe it is not good news for bond holders.

Referring to loans that would be made, post-2013, in the context of the permanent crisis management mechanism (European Stability Mechanism or ESM), the Eurogroup indicated that:

"In all cases, in order to protect taxpayers' money, and to send a clear signal to private creditors that their claims are subordinated to those of the official sector, an ESM loan will enjoy preferred creditor status, junior only to the IMF loan".

That decision raises meaningfully the expected loss on long-dated sovereign bonds, in our view.

Official support does not reduce the probability of default... If the Irish example is to be used as a reference, the fact that a government can obtain funding from its peers or the IMF does not significantly lower the fiscal hurdle and therefore may not improve considerably the market view on the probability that the government eventually defaults. Portuguese 10-year yields have hovered in a 6-7% range over the past month, so the government would not benefit greatly from funding at the same level as Ireland. Yet, Portugal would need cheap funding even more than Ireland to compensate for its lower growth potential. Spain, whose 10-year yields are currently lower, would derive no net benefit from external assistance. External assistance should therefore not alter much the market view on the size of the hypothetical loss that governments would have to impose on their creditors, should default occur - a risk that the Eurogroup recognises explicitly.

...but increases loss given default for bondholders. What resort to inter-government funding does, however, is to change the market perception of loss given default. The reason is straightforward: if official loans are senior to bondholders' claims, then any loss would have to be borne entirely by the latter. The higher the share of government borrowing that comes from the official sector, the larger the haircut would be for bond holders if default did occur.

Cheap market funding becomes less likely... The implication is that if it becomes plausible that a government has to resort at some point in the future to ESM/IMF assistance, this ought to immediately increase its cost of funding. Rising yields, in turn, raise more doubts about fiscal sustainability, undermine the government's access to markets, make it more likely that it requires assistance from its peers and validate ex post the rise in bond yields in a spiral of self-fulfilling expectations. For this reason, we think that the subordination of government bonds makes it more likely that an increasing number of governments eventually have to apply for multilateral funding support.

...especially at longer maturities. At the very least, it is increasingly unlikely that governments can lengthen the maturities of their new bond issuance, as called for by the Eurogroup. Indeed, the longer the maturity of a bond, the more likely that a substantial amount of senior official claims accumulates through the life of the bond. We project the breakdown of Irish debt between bonds and official loans, assuming that the government is entirely funded by its peers and the IMF from now on. Strictly for the sake of illustration, assume that the Irish government eventually has to impose a 25% haircut on its debt and that this haircut is entirely borne by bond holders. If this restructuring took place in 2017, the recovery rate on outstanding bonds would be 50%, reflecting the fact that half of the debt would be senior to bonds by then. If the restructuring were to take place in 2019, the recovery rate would be zero, as outstanding bonds would by then amount to no more than one quarter of the debt. Note that we are not saying that we expect Ireland or any other government to restructure its debt. Rather, our point is that the 'demotion' of government bond to subordinated status cannot but raise the price at which governments raise finance in the market and reduce their ability to do so at long maturities.

Why it matters little that restructuring is excluded until 2013. One might object that the demotion of government bonds is only partial (the Eurogroup statement only refers to ESM loans, not to loans extended under the current EFSF scheme) and that involvement of the private sector in any sovereign debt restructuring is not intended to take place before 2013. We do not find that either of these arguments provides more than partial reassurance, for the following reasons:

• Even before the European Stability Mechanism (ESM) is created, there already exists a portion of government liabilities that is senior to bonds, both in the case of Ireland and Greece and in the case of any other country that might have to resort to inter-government funding by mid-2013: IMF loans enjoy preferred creditor status;

• For a holder of subordinated claims, what matters is the amount of senior claims at the hypothetical point of default, not so much when these senior claims start accumulating. So, demotion is effective today at least for any cash flow arising post-June 2013;

• Preferred creditor status is a creature of policy rather than contract. What sovereign governments repay and to whom at a hypothetical point of default remains their decision. From this perspective, the Eurogroup could hardly have been more explicit on policy intentions regarding where private creditors stand relative to official ones.

It is all government bonds that have been downgraded. One might also object that subordination only takes place if and when a country has to resort to peer support, so that demotion of government bonds applies only to Greece and Ireland. We disagree.

Subordination is not only relevant when a government has already required external support, but also if it may do so in the future. And as we have had the opportunity to argue in previous months, almost all governments of advanced economies are deep in negative equity. The question is no longer whether Spain is different from Portugal, Portugal different from Ireland and Ireland different from Greece, as repeated (with admittedly valid arguments) by European leaders over the past weeks. The fact is that it is the entire asset class of government bonds that has been downgraded at once. Contagion does not trickle from country to country and government to government. It affects all at once, albeit at different degrees.

As Joachim Fels and Elga Bartsch have underlined, the Eurogroup decision institutionalises and extends a situation that was already reflected in market pricing and behaviour (see The Global Monetary Analyst: On the Question of QECB, December 1, 2010, and also Laurence Mutkin's When Is a Government Bond Not a Government Bond? September 22, 2010). Peripheral European governments have lost - or are close to losing - their investor base and are unlikely to regain it until debt sustainability has materially improved, one way or the other: one way means default, the other means a necessarily slow fiscal consolidation. In our view, however, the impact of the Eurogroup's announcement goes further than the Euroland periphery. Or perhaps more exactly, it blurs the boundary between the periphery and the core and makes the investor base of more governments more fragile - as seen by the widening of Italian and Belgian sovereign spreads. 

Meanwhile, senior bank debt remains protected - so far. In this context, it is ironic that a similar wholesale demotion had been carefully avoided as regards senior bank debt in the negotiations over the terms of the Irish bailout. The decision to protect senior creditors of Irish banks from default was fully justified, in our view, by the systemic implications that it would likely have generated. Default on senior bank debt in one case would have caused investors to reassess the risk profile of most if not all senior bank debt across Europe, with three likely implications: a reduction in the availability of wholesale funding (causing bank deleveraging to proceed at an accelerated pace), a rise in the cost of funding (mirrored by a rise in the cost of credit) and a shortening of the duration of bank liabilities (leading in turn to a reduction in the availability of funding for long-term productive investment). All of this could have pushed Europe back into a genuine credit crunch.

Debt jubilee or debt assumption? If the aversion of the market to funding governments through what appears to be subordinated instruments spreads beyond a few small peripheral countries, then one of two things could happen: either much more radical solutions are devised to keep all governments funded at sensible rates (i.e., those governments that have kept access to funding lend this access to their peers, as they did for their banks in autumn 2008); or the likelihood of a wide-scale debt restructuring rises quickly.

This is not to say that the decision to split government liabilities between senior and subordinated is misguided. On the contrary, we believe that it can lead to a credible and lasting solution to the sovereign debt crisis in Europe - and beyond.

Why splitting government debt between senior and subordinated debt is sensible... Put simply, if a government's solvency is seriously in question, the only way it is going to be financed - at least at affordable rates and without implicit subsidies - is by giving new creditors additional protection relative to that enjoyed by existing ones. This is not only sensible but also standard practice: when the IMF extends a loan to a government, it benefits, as mentioned above, from preferred creditor status. IMF involvement automatically deteriorates the position of other creditors. The same applies to any central bank acting in its capacity of lender of last resort to a bank: so long as central bank loans are collateralised, this necessarily deteriorates the position of the bank's unsecured creditors.

...and why it is desirable. The benefits of splitting sovereign debt between various levels of seniority go further, in our view, than facilitating cheap funding. One strong argument to avoid government default is that government bonds are 'special'. They are a safe haven in times of stress; a reference for pricing of financial contracts; preferred collateral; core and comparatively safe assets for banks, insurance companies and pension funds; etc. This creates a well-known dilemma between moral hazard and systemic risk: simplifying the argument somewhat, if governments are over-indebted and default is to be avoided, then either of two equally undesirable outcomes has to materialise; either inflation has to rise (where the central bank can monetise government debt) or fiscal transfers have to take place (in a monetary union where the central bank refuses to monetise government debt). That chain of argument, however, relies on the fact that all government debt is pari passu and that governments cannot fail 'by degree'. It falls apart if government debt were to be split into two parts: a senior part, constrained by a rigid debt ceiling, which alone performs the roles listed above and is essentially a true risk-free asset; and a subordinated part, on which governments can default without prejudice to the former part and therefore without systemic implications. The benefits of splitting government debt between a senior and a junior part in the euro area have been argued notably by Jacques Delpla and Jakob von Weizsäcker or Mario Monti. 

Killing two birds with one stone: how an assumption scheme would operate. In a context where the Eurogroup has effectively endorsed the concept of subordination of government debt, we think this makes it possible at the same time to avoid a debt jubilee and to strengthen the European fiscal framework. The debate on assumption has now started in earnest as some senior politicians have expressed support for common bond issuance among European governments. In our view, the best way to do so would be the following:

• Create a European agency that issues common bonds on behalf of governments, with joint and several guarantee by all euro area member states.

• The monies borrowed in the market would be on-lent to each government according to their borrowing requirements, at no extra cost - i.e., the agency would pass on its (low) borrowing costs equally to all governments;

• The amounts lent to each government would have to be strictly consistent with what is required under a credible but rigid fiscal adjustment programme, bringing first the deficit/GDP ratio to 3% of GDP by 2013 and thereafter bringing the debt/GDP ratio to 60% of GDP by a (distant) date, identical for all governments. This would imply a more demanding and sustained fiscal effort for those countries whose initial debt level is furthest away from the reference value.

• The claims of the agency on each individual government would be, consistent with the wording of the Eurogroup statement of November 28, 2010, senior to any other claim (barring any outstanding IMF loans).

• Any government borrowing in excess of the amounts implied above would take place in its own name and would by construction be explicitly subordinated debt, therefore implying a higher degree of risk for the buyers.

• The amounts that the agency can raise and on-lend to governments could not under any circumstance be altered unless by unanimous decision of all member states.

• Recognising the senior nature of the agency claims and the riskier nature of any subordinated debt, regulatory capital relief should realistically only apply to holdings of the common debt. Arguably, after a transition period, the ECB may choose to only accept as collateral the common senior debt, at the exclusion of subordinated securities issued by individual governments.

Four benefits of an assumption scheme. The merits and implications of setting a scheme in this exact form would be, in our view:

• It would ensure that all governments can remain funded at affordable rates. The credit quality of commonly issued bonds would benefit from a double protection: i) the joint and several guarantee of all governments; and ii) the fact that the common issuer benefits from a senior claim on government resources - for amounts that by construction never exceed the repayment capacity of any individual government. That second point would, incidentally, also provide protection to peer governments in their capacity as guarantors (explicitly: Germany would not need to be concerned with a default by Ireland since it would only ever guarantee an amount of Irish debt that Ireland can repay and will repay before any other).

• It would create a financial incentive for fiscal prudence and abidance to the rules of the Union, by raising materially the cost for any government of accruing debt in excess of the agreed trajectory. The yield that governments would need to pay on issuance of junior debt is uncertain but would likely be much higher than that of the common bonds.

• It would reduce moral hazard and limit the systemic consequences of sovereign default, by removing the fungibility between debt issued at 'safe' levels and debt issued above these levels. Indeed, default on the subordinated part of the debt need not have any systemic implication, since this risky fraction of the debt would not be likely to be held in any significant amount by banks, insurance companies or pension funds, nor would it be used as collateral for financial transactions.

• It would crystallise the ongoing shift from peer pressure to peer control in the European multilateral fiscal surveillance framework and place that control in the hands of those governments - such as Germany - most supportive of fiscal prudence. That control derives from the veto that each government maintains over a loosening of the rules applying to all.

There is to our knowledge no current plan to go down the route described above. But we believe that the ongoing acceleration of the pace of the crisis means that this and alternative schemes are increasingly likely to be seriously considered and debated.

Good for governments. But good for bond holders? Less so. We believe that an assumption scheme such as that described above would meet both the needs of the weakest governments (which would benefit from continuous, affordable funding) and the requirements of the more robust governments (which would see their influence reinforced and would avoid fiscal transfers). However, it would only improve marginally the position of current bond holders. The improvement would come from the fact that cheaper funding on new debt would lower the fiscal hurdle for the most challenged governments - unlike in the case of Ireland - and thereby improve the likelihood that governments eventually repay all their financial debt. It would only be marginal because, at the same time, an assumption plan would mean that outstanding bonds would become 'legacy bonds' and see their status as subordinated debt confirmed, no longer only de facto, but now also de jure.

Whichever route Europe might choose to follow - debt assumption or jubilee - currently outstanding bonds of at least peripheral European countries should therefore be seen as subordinated instruments, and are likely to remain valued accordingly by the market.