The latest and greatest plan to bail out Europe revolves around using the recently expanded and ratified €440 billion EFSF, and converting it into a "first loss" insurance policy (proposed by Pimco parent Allianz which itself may be in some serious need of shorting - the full analysis via Credit Sights shortly) in which the CDO would use its unfunded portion (net of already subscribed commitments) which amount to roughly €310 billion, and use this capital as a 20% "first-loss" off-balance sheet, contingent liability guarantee to co-invest alongside new capital in new Italian and Spanish bond issuance (where the problem is supposedly one of "liquidity" not "solvency"). In the process, the ECB remains as an arm-length entity which satisfies the Germans, as it purportedly means that the possibility of rampant runaway inflation is eliminated as no actual bad debt would encumber the asset side of the ECB. A 20% first loss piece implies the total notional of the €310 billion in free capital can be leveraged to a total of €1.55 trillion. So far so good: after all, as noted Euro-supporter Willem Buiter points out in a just released piece titled "Can Sovereign Debt Insurance by the EFSF be the "Big Bazooka" that Saves the Euro?" there is only €900 billion in financing needs for the two countries until Q2 2013. As such the EFSF would take care of Europe's issues for at least 2 years, or so the thinking goes. There are two major problems with this math however, and Buiter makes them all too clear.
One: rating downgrades and ongoing deterioration - should the financing needs of not only Spain and Italy, but also Belgium and France, post its inevitable AAA-downgrade, need to be funded the total insurable amount rises to €2,371.6 billion through Q2 2014. And since there needs to be headroom, and since a number of €3 trillion has been thrown around, there is just no practicable math of how one gets from the current committed funding to the €726 billion that would be required for the full funding amount, especially with a AAA-rating still retained by the CDO. Second, and just as important, is taht the 20% first loss ratio "may well be far too optimistic." Simply said, a far more realistic recovery rate would be one of 50-60% meaning a first loss guarantee of 40-50% will be required, which collapses the total insurable "pot" to about €600 billion. Buiter's unpleasant, for Allianz, Merkel and Sarkozy conclusion is that "that would likely not fund the Spanish and Italian sovereigns until the end of 2012. It would not be a big bazooka but a small pea shooter."
So just like we proclaimed the second Greek bailout DOA when it was announced, so we proceed to say that, once the market has had the time to digest what is really happening and proceeds to go after the weakest links in the plan one by one, that the EFSF is also dead on arrival. But in the meantime, it will buy the Eurozone a little more time to pretend that all is well, and provide skeptics with very attractive short-term EURUSD shorting abilities.
Here are some of the key observations from Buiter. First, he estimates what the pure unencumbered capital of the EFSF could be in an ideal case, starting with the flawed number, previously cited by Aliazn, of the full €780 billion in Guarantee Committment which can allegedly be used and leveraged 5-fold to get to the critical €3+ trillion number.
Total guarantee commitments from the 17 EA member states are just short of €780bn (see Figure 1).
From that number, we here subtract i) guarantees by the current set of countries with ‘stepping-out’ status (all of which are out of the primary sovereign debt markets for the time being), ii) guarantees provided by Italy and Spain, as these two countries are not credible guarantors because they are highly likely to default themselves (on their own sovereign debt and on the guarantees they provide to the EFSF) whenever there is a call on the guarantees provided by EFSF to any of the non-stepping out sovereigns, iii) other countries that, like Spain and Italy today, could potentially benefit in the future from EFSF guarantees of their sovereign debt, or are likely to ask for ‘stepping-out’ status, iv) existing and likely imminent EFSF commitments.
Next, he removes the "step out" guarantors sequentially:
Greece (€29.1bn of notional EFSF guarantee commitments), Ireland (€12.4bn) and Portugal (€19.5bn) have become “stepping-out guarantors”, that is, their guarantees cannot be called upon as long as they remain Troika programme countries receiving funding from the EFSF. Portugal remains liable as guarantor in respect of notes issued prior to the time it became a stepping-out guarantor. Estonia (€2.0bn) is only a guarantor in respect of notes issued after the effective date of the Amendments to the EFSF Framework. This means that, as of today, the aggregate of the active guarantee commitments for the guarantors which are not stepping-out guarantors is € 726bn
Then, Italy and Spain:
But, in addition to the guarantors belonging formally to the stepping-out guarantors category, there are the de-facto stepping-out guarantors which currently need the assistance of the ECB, through its Securities Markets Programme of outright purchases of sovereign debt in the secondary markets, to secure funding on affordable terms. Spain (€92.5bn of EFSF notional guarantees) and Italy (€139.3bn) are in that category since the ECB resumed its SMP purchases at the beginning of August 2011. Indeed, the insurance programmes proposed by Achleitner and Kapoor are mainly aimed at ring-fencing the Spanish and Italian sovereigns – and, lurking behind them, the sovereigns of Belgium and France - and ensuring continued access to the funding markets for them. Again, Spain and Italy cannot insure themselves when they are both at clear and present risk of being cut off from market funding at affordable interest rates and are therefore unlikely to make good on any notional guarantees they have offered to the EFSF, should the EFSF have to call on these guarantees. Taking Spain and Italy out of the €726bn guarantee pot would reduce it to €494bn
Next, remove pre-existing commitments.
Under the Troika programme for Portugal, the EFSF has committed €26bn, of which €5.8bn has been disbursed thus far. The commitment of the EFSF to the Irish Troika programme is €17.7bn, of which €3.3bn has been disbursed thus far. Subtracting the commitments of the EFSF to the Portuguese and Irish Troika programmes leaves €445.5bn uncommitted.
Under the proposed €109bn second Greek programme, the EU/EA contribution is likely to be at least two thirds, or €72.6bn, if the ‘two thirds for the EU/EU, one third for the IMF’ division of costs for the Greek Loan Facility and for the other Troika programmes is maintained. The European contribution could be higher because, as of now, the IMF has not given any formal commitment that it will make a financial contribution to the second Greek programme. The European contribution could be lower if the IMF continues to take on one third of the total official funding commitment and if the revision of the terms of the 2nd Greek bailout facility results in a smaller total official funding contribution than the €109bn announced on July 21.
Assuming, for the moment, that the two potential sources of variation cancel out, the EFSF commitment would be two thirds of the original agreement or €72.6bn. Of this €72.6bn, at most €11.5 billion could be funded by the European Financial Stabilisation Mechanism (EFSM), the supranational source of funding backed by the EU budget. This is because only €11.5bn remains uncommitted of the €60bn EFSM facility. So €61.1bn has to be subtracted from the total available for sovereign debt insurance, leaving €384.4bn uncommitted. If, as seems likely, the EFSM does not contribute to the second Greek programme, the sovereign debt insurance pot would go down to €373.9bn.2 If the IMF decides not to co-fund the second Greek programme, it goes down to €337.1bn. In addition, roughly €27bn of the European contribution to the Greek Loan Facility that has funded the first Greek bailout programme is still to be disbursed. Given the difficulties and funding rates that some of the EA creditors face in raising funding for the Greek Loan Facility (which is funded, unlike the EFSF, on a bilateral basis), it is at least plausible that the remaining tranches of the first Greek programme (i.e. €27bn) will be paid out of the EFSF pot. This reduces the EFSF resources to either €346.9bn or €310.1bn, depending on whether the IMF co-funds the second Greek programme.
Sovereign debt insurance is not the only remaining claim on these resources. The revised EFSF framework agreement mentions explicitly the possibility that the EFSF offers support for recapitalising euro area banks, including those in non-programme countries, though any such assistance would still need to be routed through the respective sovereigns.
The EFSF has already contributed to the recapitalisation of banks in programme countries (Ireland and Portugal) through loans to their governments. Sovereign debt insurance is not even mentioned explicitly under the intended uses of EFSF resources. No doubt, however, it could be shoehorned in under “precautionary facilities”. A very conservative estimate for the amount the EFSF ought to set aside for the recapitalisation of financial institutions in non-programme countries (in the first instance Spain and Italy, and beyond them possibly Belgium, France and other core EA member states) would be at least €50bn, leaving at most €296.9bn and possibly just €260.1bn for sovereign debt insurance. Clearly, banking sector recapitalisation outlays could be much larger. For instance, Citi’s equity research Banks team estimates that banks from EA countries alone that have difficulty to access private capital markets (Greece, Italy, Ireland, Portugal and Spain) currently would need €104bn to bring their Core Tier 1 capital ratio to 9%.
These estimates don’t allow for the possibility that Belgium, now on negative outlook for all three rating agencies (see Figure 1) and with historically high sovereign 5-years CDS spreads and spreads over 10-years Bunds, might join the ranks of countries needing sovereign bond insurance rather than contributing to the resources needed to provide such insurance. This would reduce the sovereign debt insurance pot to €269.9bn or €233.1bn. Having France move from the insurer to the insured category would deplete the resources available to €111.4bn or €74.6bn.
Naturally an insurance fund working with just €74.6 billion, regardless of the amount of first loss assumption is a joke. So for all intents and purposes, Buiter has used the €310 billion number bolded above.
What does this mean for total maximum notional insurable?
[There is] €310bn if France remains among the insured and no banking sector support is provided by the EFSF, around €260bn with a €50bn provision for banking sector support, and less than half that if France were to join the insured.
With a potential first loss guarantee of just 10%, the resulting maximum amount of new debt issuance that could benefit from a guarantee could be up to €3.1trn without EFSF support for bank recapitalisation or, in the more likely case of a €50bn bank recapitalisation contribution, up to €2.6trn. A 20% first loss guarantee (a figure that seems to be in the air quite a bit) would imply maximum issuance amounts of around €1.55trn or €1.3trn, respectively, and a 40% first loss guarantee would result in maximum issuance amounts of around €777bn or €650bn.
So what is the fundamental reasoning for a bailout mechanism? Why to make sure that the trillions in European debt that has to be refinance and rolled over the next 3 years, are, respecitvely, either refinance or rolled. Let's take a look at what amount we are talking about here. First the stock amount, which means insuring existing debt.
With the insurance approach, the EFSF can target flows of new sovereign debt issuance in the primary markets while leaving the outstanding stocks of sovereign debt uninsured. Under current circumstances, the EFSF could focus fully on guaranteeing new debt issues by Spain and Italy, the two sovereigns that are still in the markets but at risk of being frozen out of the markets through a fear-driven denial of market funding by the private sector. Greece, Ireland and Portugal have been taken out of the market and are being funded (in part also through the EFSF) through loans.
The ability to insure just the new flows rather than both the new flows and the outstanding stocks is valuable, as (see Figure 2) the outstanding stocks would swamp the capacity of the insurance facility.
Italy and Spain together have just under €2.5 trillion worth of general government debt outstanding. Tradable Spanish and Italian sovereign debt alone amounts to €2.1 trillion. Adding Greece, Ireland and Portugal raises general government debt to €3.1 trillion and tradable government debt to €2.6 trillion. Adding Belgium would raise these totals to €3.5 trillion and €2.9 trillion. In the perhaps unlikely case that France would need sovereign debt insurance, targeting the stocks rather than the flows would require taking care of €5.1 trillion of gross sovereign debt or €4.3 trillion of tradable government debt.
These numbers are beyond the size of even the most optimistic estimates of the most audacious of rescue umbrellas. Fortunately, to avert a funding disaster for the vulnerable sovereigns, only the flows of new funding need to be insured. As Figure 3 makes clear, these flows, while large, are more manageable than the stocks.
Flows simply means focusing on new issuance, to guarantee there is a natural market bid from the private sector, combined with a helping hand from the EFSF. How much debt would have to be insured? Well, depend on which of three scenarios we are looking at: i) just Spain+Italy; ii) Spain+Italy+recently imploding Beligum; or iii) Spain+Italy+recently imploding Beligum and most recently collapsing France (take one look at the OAT-Bund spread to see what we are talking about). The numbers are not any prettier.
Focusing on the flows that matter, the total financing needs of Spain and Italy, Figure 3 gives us €153bn for 2011 Q4, €557bn for 2012, €348bn for 2013 and €295bn for 2014. For the seven quarters from 2011 Q4 to 2013 Q2 (the last quarter before the start of the European Stabilisation Mechanism, the successor of the EFSF) the total financing needs of Spain and Italy are €881bn.5 For the 11 quarters from 2011 Q4 to 2014 Q2, they amount to €1,196bn. Adding Belgium raises the seven-quarter funding total to €1,006bn and the 11-quarter total to €1,377bn. This would be manageable if the markets were willing to fund these amounts with a 20 percent first loss guarantee, assuming the triple-A insurance capacity of the EFSF is around €300bn and the market provides new funding to the sovereigns at acceptable rates with a 20 percent first-loss guarantee. If, however, France were to join the ranks of the countries needing external official guarantees to fund themselves, the seven-quarter total funding need would go up to €1,684bn and the 11-quarter total to €2,372bn. Neither would be manageable with the existing size of the EFSF resources.
Here Buiter makes a great point: the EFSF would effectively create two markets in each nation's sovereign securities: existing, or ex-guarantee, and post EFSF, including the guarantee:
Of course, focusing the insurance on the flows of new funding alone leaves the prices of the outstanding stocks of sovereign debt to be determined in the secondary markets, with the ECB most likely absent from the secondary markets if the insurance option is implemented. We see no point in supporting an orderly secondary market through outright purchases of sovereign debt under the Securities Markets Programme, even if the price in the (uninsured) secondary market is significantly below that in the (insured) primary market. Banks and other systemically important financial institutions that have to mark their holdings of sovereign debt to market will of course be adversely affected by any wedge between the primary and secondary market prices, and may even have to raise additional capital to deal with any mark-to-market losses, but that is the way of markets and market economies.
But wait there is more. The biggest weakness of the EFSF is the embedded assumption for how much first-loss a potential investor will be ok with. According to Allianz and the Europeans 20% should be sufficient. It won't be.
Even in the absence of a panic, the (average) 20 percent first-loss ratio assumed in much of the calculations in this note may well be far too optimistic.
The historical recovery rate for sovereign defaults from 1983 up to 2010 reported in Moody’s (2010) is only 53 percent (issuer weighted) and as little as 31 percent value-weighted. Sturzenegger and Zettelmeyer (2005) find, in a study covering sovereign defaults between 1998 and 2005, haircuts ranging from 13 percent to 73 percent. Clearly, the cost of default to the defaulting sovereign has an element of fixed cost in it. The reputational loss associated with a breach of contract does not double if the recovery rate falls from 80 percent to 60 percent. Even if sovereign defaults are unlikely and infrequent, the size of the NPV loss for the investors conditional on a default having occurred is therefore likely to be large: if the sovereign is going to default at all, she might as well be hung for a sheep as for a lamb.
And any incremental rise in the first loss guarantee threshold implicitly removes the leveragability of the underlying notional. Said otherwise, a €310 billion pot which insures 50% of losses, means a mere €620 billion in notional can be guaranteed: a failure off the bat which the market will stampede over.
Which brings us to Buiter's less than glowing conclusion:
The partial/first loss sovereign debt insurance proposal has merits. Its principal value is that it permits the decoupling of the essential flow funding aspect of the Euro Area sovereign debt crisis from the much less important stock valuation aspect. Insurance can be offered for new issuance of sovereign debt in the primary markets without extending the offer to the outstanding stock of debt – the debt traded in the secondary markets. The further option of using the partial insurance route to enhance sovereign debt issued as part of a sovereign debt restructuring, e.g. through a ‘voluntary’ exchange along the lines of the Greek PSI proposal, is also valuable.
There are three problems with the specific proposal made by Achleitner and Allianz. First, we believe the arithmetic materially overstates the total amount of debt issuance that could be insured with the existing resources of the EFSF. To get to the €3 trillion worth of debt touted in some of the proposals, we estimate the EFSF would, with an average first-loss guarantee of 20 percent, have to make available for its insurance activities nearly all of its notional €726bn worth of non-stepping-out member state guarantees, and these resources would have to be viewed by wouldbe insurance purchasers as being of triple-A quality. This would mean (1) that the EFSF would have to renege on all its outstanding and pending commitments other than insurance, (2) that Spain and Italy would effectively be insuring themselves, and (3) that a triple-A guarantee capacity of €440bn is miraculously transformed into one of €726bn.
Second, the insurance mechanism is not fool-proof or disaster-proof. There is no guarantee that, in a panic, the most generous terms on which the insurance could be offered (say, for free) would be attractive enough to bring in sufficient private buyers of the insured sovereign debt. Failed auctions and sovereign default are not ruled out. A standby purchaser of last resort for sovereign debt is required. This standby purchaser of last resort would have to be either an official entity or a private entity created, funded and directed by the official sector.
It could be the ECB through the SMP along the lines of its interventions buying Greek, Irish and Portuguese sovereign debt since May 2010 and Spanish and Italian sovereign debt since August 2011. However, because of the Treaty prohibition on the central bank funding sovereigns directly, the ECB can purchase sovereign debt only in the secondary markets, which would be an inefficient use of public resources. The ECB could end up owning much of the outstanding stock of periphery EA debt to achieve a relatively limited amount of new funding by the periphery sovereigns.
It could be the EFSF, which can operate in the primary issue market, even without the EFSF being granted eligible counterparty or ‘bank’ status allowing it to borrow from the Eurosystem against collateral. That would, however, restrict the volume of such purchases to the uncommitted part of the EFSF’s own resources. To have a bigger standby bazooka, either the EFSF, or some special purpose vehicle created by the EFSF (possibly in conjunction with the European Investment Bank, which has eligible counterparty status with the Eurosystem) would have to be granted eligible counterparty status for collateralised borrowing from the Eurosystem. The remaining resources of the EFSF could be used to provide capital for this ‘Bazooka Bank’, and/or to guarantee the loans from the ECB to the Bazooka Bank, which would in any case be secured with the sovereign debt purchased in the primary market by the Bazooka Bank.
Third, and last, is the abovementioned shortfall in the first-loss tranche, which will need to be substantially increased, potentially doubled, to provide bidder with a safety of mind that even in a worst case, read 60% recovery scenario, they will not suffer catastrophic losses.
We therefore are sceptical that, if there is a reasonable expectation that the recovery rate following a sovereign default in the Euro Area could be as little as 60 percent or 50 percent, that the markets would be happy to fund these sovereigns at sustainable interest rates to the sovereigns, with just a 20 percent first-loss rate, even if this insurance were granted free of charge. A 40 or even 50 percent first-loss rate might well be required. And that would reduce the amount of new issuance that could be funded with an EFSF insurance pot of, say, €300 bn at most to just €750bn or even €600bn. That would likely not fund the Spanish and Italian sovereigns until the end of 2012. It would not be a big bazooka but a small pea shooter.
As we said: once the market (forget the polticiians: they would want nothing more than for nobody to see this analysis) is made comfortable with this actual math, it will realize that the EFSF as an Allianz rescue facility, pardon, insurance fund, is Dead on Arrival. After all, and logically, if this formulation was the best and safest one, it would have been proposed months, if not years earlier, not been used as the last ditch Deus Ex Machina, with just 5 days until the European Summit.
We can't wait until this latest episode of cognitive bias (EFSF will work) clashes with the hard reality of math and numbers.