Greek CDS Exposure: Proving the ISDA and Bloomberg Wrong

According to numerous news reports, the net exposure arising as a result of Friday’s decision to declare the activation of Greek CACs a credit-default event amounts to approximately $3.16 billion (assuming that the bonds achieve no value at the ISDA auction on March 19).

Where is the data coming from?

On the same day that the ruling announcement was made, the ISDA issued a FAQ document relating to the Greek Sovereign CDS credit event. Both these documents can be downloaded at http://www2.isda.org/greek-sovereign-cds/. This document makes reference to the $3.16 billion net notional as disclosed by the DTCC (The Depository Trust and Clearing Corporation). In fact, the publication erroneously links to Table 5 of the DTCC data – the actual figure is found in Table 6. Link: http://www.dtcc.com/products/derivserv/data_table_i.php?tbid=6. Table 6 discloses the Top 1000 Reference entities for the swap arrangements (ie. the issuers of the instruments being covered by the credit default swaps). Under the heading “Hellenic Republic”, the sovereign debt gross notional is disclosed as $68,901,331,331; and the net notional is disclosed as $3,162,579,257.

What is the problem that people are having/should be having?

The problem appears to be twofold:

1. The preliminary list of bonds submitted to the ISDA for restructuring credit event consideration included issuers other than the Hellenic Republic. 
2. Is the sovereign debt exposure figure presented by the DTCC a fair reflection of economic reality? 

Problem 1: The Other Issuers on the Preliminary List

From my reading of the ISDA public announcements, the problem is as follows:

1. The preliminary list of bonds submitted for review (released by the ISDA on March 7 – this can also be downloaded from the link identified above) includes issuers other than the Hellenic Republic. No basis for the inclusion is given, nor is it required. The list is not final – rather, these are the bonds that the public has asked the ISDA to consider. 
2. It is suggested that the reason for inclusion is on the basis that many of these issues have guarantees issued by the Hellenic Republic. Given that the trigger event was the activation of CACs, one could infer that the default will only extend to those bond issues where CACs have been retroactively imposed (ie. only Sovereign Debt). 
3. The ISDA has not identified the bond issues to be affected by the default announcement: the wording is specific to the occurrence of a Restructuring Credit Event. 
4. According to the announcement, the EMEA Determinations Committee is currently in the process of reviewing the preliminary list of submitted bonds. 
5. Therefore, the magnitude of the default event will only be set after the EMEA DC formally identifies the bonds affected. 

Can impact of the other issuers be quantified?

The issuers identified in the preliminary list of submitted bonds are:

1. The Hellenic Republic 
2. Athens Urban Transportation Organisation (OASA) 
3. Attica Bank S.A. 
4. EFG Eurobank Ergasias S.A. 
5. Hellenic Railways (OSE – Organismos Sidirodromon Ellados) 
6. National Bank of Greece S.A. 
7. New Economy Development Fund S.A. (Taneo) 
8. Proton Bank S.A. 
9. Aeolos S.A. 
10. Ariadne S.A. 
11. Agricultural Bank of Greece S.A. 
12. Alpha Bank AE 
13. Piraeus Bank S.A. 

According to Mark J. Grant in his article “The Eight Hundred Pound Greek Gorilla Enters The Room” (link: http://www.zerohedge.com/news/eight-hundred-pound-greek-gorilla-enters-room) – these other bond issues are going to increase the CDS impact number up to $79 billion.

I’m not sure where that number comes from, but I’m going to assume (for this section) that the DTCC figures are correct. If they are, then with the exception of the Hellenic Republic, none of the above issuers feature in the Table 6 Top 1000 reference entities. The smallest of the Top 1000 (by net notional exposure) is Nextel Communications, with a Gross Notional of $1,649,302,354 and a Net Notional of $105,475,330. By inference, therefore, the maximum additional net notional exposure to be added by the other 12 entities identified is an extra $1.3 billion ($105 million by 12).

However, as I have previously stated, this is unlikely to be a real consideration as the triggering event is the activation of the Collective Action Clauses – an event which should have no direct default impact on government guarantees given to state-owned entity and domestic bank issues.

Problem 2: Is the DTCC’s calculation correct?
Renowned investor Jim Sinclair (JS) in an interview with King World News suggests that the real question to be asked should be: is the amount calculated by the DTCC correct? He says that the DTCC amounts calculated are only in respect of transactions taking place within the US (ie. they exclude foreign bank transactions, as well as transactions undertaken by the foreign subsidiaries of US banks). If that were the case, then we would have to look for the real information in the data presented by the Bank of International Settlements (BIS).

The latest data available from the BIS takes us up to June 2011. According to Jim Sinclair, the total CDS market approximates $37 trillion, of which over 50% must relate to Greek debt. The first figure is supposedly drawn from BIS statistics; the second is based on his 50 years of market experience.

Well, the first figure is factually incorrect (or perhaps chronologically incorrect – it’s always possible that the number referred to a different point in time). At the end of June 2011, the total outstanding CDS market, according to BIS, was $32.4 trillion (net exposure).

As regards the second, the BIS gives a breakdown of the CDS market based on reference entity sector. Of the $32.4 trillion, only $2.9 trillion actually relates to Sovereign Debt. The only other potential impact is in the sector of the CDS market held over CDO instruments, which may or may not include Greek sovereign debt issuances in default. This sector of the market constitutes a net notional $0.9 trillion. So in total: a potential exposure of $3.8 trillion.

Obviously, this is significantly larger than the $3.16 billion being used by the ISDA (over 1000 times larger). And Jim Sinclair does make a good point: the DTCC tagline for its data is that “this section provides comprehensive reports on the vast majority of CDS contracts registered in the Warehouse’s global repository” (link: http://www.dtcc.com/products/derivserv/data/index.php). Obviously, CDS contracts that are not registered with the DTCC would not be included in this data. And it certainly sounds like the DTCC captures only the US portion of the market.

As a reasonability check, compare the total of the Table 6 Top 1000 reference entities to that of the total CDS market according to the BIS figures. The total net exposure of the DTCC Top 1000 is only $1.1 trillion, compared to the $32.4 trillion of the BIS. If we say that the DTCC figures approximate the BIS figures, the balance of $31.3 trillion must then be made in reference to other issuers. Given that the Top 1000 include governments and all major corporate and parastatal bond issuers – this implies that the majority of the CDS market has been taken out against small issuers! Highly unlikely.

As a final check, one should consider what the figure of $3.16billion is implying about the global debt market. I have used Table 15B of the BIS quarterly statistics (link: http://www.bis.org/publ/qtrpdf/r_qa1203.pdf), which breaks down the Total International Bonds and Notes by nationality of issuer. Whilst this statistic uses total international bonds and notes outstanding rather than just sovereign debt, or a combination of both international and domestic debt: the point is to illustrate the proportion of the international debt market that is made up by Greek debt. 

Total International Bonds and Notes Outstanding (in US billions)
Greece	397.40
The World	27 819.70
Proportion represented by Greece	1.43%
Total CDS Net Notional Exposure (in US Billions)
CDS Net Notional Exposure (DTCC)	3.16
World Net Notional Exposure (Sovereign Debt) - from BIS	2 907.80
Proportion represented by Greece	0.11%

If we just assume that all investors were blind and equally distributed their CDS investments between countries, we would expect Greece’s net notional exposure to approximate its portion of the debt market. Even more so, we would have expected investors to flock to CDS instruments over Greek sovereign debt, given the economic fundamentals of Greece’s position (ie. you’re much more likely to buy insurance when your house is made of untreated wood and situated next to a blacksmith, than if your house is made of stone). Instead, the ISDA (and Bloomberg) would have us believe the opposite, that investors underweighted their CDS exposure to Greek sovereign debt.

Is there a reasonable explanation for underweighting?

I think that there is an argument to be made here. After all, why has a Greek default been delayed for so long? A year ago (even 6 months ago), the Euro leaders were all against causing a credit-restructuring event for fear of triggering CDS instruments and causing contagion. And now suddenly, everyone is keen for a default. Including Greece, who elected to action the CACs to top up the extra 14% of the debt – where they could have just restructured the 86% voluntary portion and gotten away with it. So the question is: why now?

I think that the delay has accomplished two things:

1. If I were speculating on Greek sovereign debt defaults three or four years ago (back when CDS instruments were cheap), I would have purchased CDS instruments over the debts issues that were falling due in the short-term – as these would have been much more likely to default (as their repayments would have fallen due sooner). By stretching out the time to default, some of the original CDS instruments will have lapsed, as Greece has managed to meet its debt obligations up and until this point; albeit with the assistance of the first bailout. And in the interim, as the market realised the Greek crisis, the spreads on Greek Sovereign CDS widened making it almost prohibitively expensive (and therefore unprofitable) to purchase more CDS instruments. Greece and the Eurozone may have faked it until they made it. 
2. The delay also gave the major banks the chance to recapitalise in anticipation of a default, and to rebalance their positions.

So perhaps the ratios sound about right.

Conclusion
Either way, the $3.16 billion figure is wrong. If anything, it is only the lower end of the range. What we can say is that there is definitely $3.16 billion of net exposure – but that it seems to only represent the US exposure. The rest of the world’s net exposure is floating somewhere in the $2.9 trillion of the BIS statistics.