Observation #1. Default Insurance Matters.
First, the BIS data very helpfully breaks exposures into two pieces: direct exposures, which basically means creditors who own bonds issued by one of the PIGs; and indirect exposures, which for the most part means agents who sold default insurance to creditors, primarily through credit default swaps. As summarized in the following table, it seems that approximately 30% of total potential exposures to debt from the PIGs are covered by default insurance (see the figures in red). Put another way, if one of the PIGs defaults, creditors who actually hold bonds from that country will absorb about 70% of the losses, while agents (primarily banks and insurance companies) that sold insurance against the possibility of default will have to cover the remaining 30%. That's not a trivial amount. (All figures below are in billions of USD, as of the end of 2010.)
Observation #2. Direct Exposure in Europe, Indirect in the US.
The table above also hints at striking differences between how European and US creditors would be hit in the case of default by one of the PIGs. If Greece were to default, for example, approximately 94% of the direct losses would fall on European creditors, and only 5% would fall on US creditors. However, US banks and insurance companies would have to make about 56% of the default insurance payouts triggered by such an event, while European agents would make only 43% of those payouts.
The next table illustrates this difference even more starkly. In the case of Greece and Portugal, the vast majority of the losses that would be borne by creditors in Europe would be direct losses. In fact, French and German creditors would almost certainly be substantial net recipients of default insurance payments. (That's less clear in the case of Ireland.) Meanwhile, US financial institutions would have to make substantial net default insurance payments, which would account for between 80% and 90% of all losses borne by the US in the case of default (see the figures in red below).
Observation #3. Similar Overall Exposures in Europe and the US.
Finally, it's worth noting that once you account for the substantial payouts that US agents will have to make to European creditors in the case of a default by one of the PIGs, financial institutions in the US have roughly as much to lose from default as those in France and Germany. (See the figures in blue in the table above.) The apparent eagerness of US banks and insurance companies to sell default insurance to European creditors means that they will now have to substantially share in the pain inflicted by a PIG default.
This has some important implications. First, US and European financial institutions are likely to have very different incentives as negotiations regarding debt restructuring and reprofiling proceed. US banks and insurance companies are surely delighted with the "soft restructuring" that is currently being discussed. Such a partial default would probably not trigger default insurance payments, and so the pain would be borne almost exclusively by European institutions. On the other hand, some time soon it seems likely that European creditors will begin to prefer a "hard restructuring" that would require default insurance payouts from the US institutions that sold such insurance. Given how strikingly one-sided the net default insurance payments will be (from the US to Europe), it's easy to imagine how that could shape future negotiations over debt relief for the PIGs.
Second, there's an interesting puzzle here. Why have European and American financial institutions behaved so differently when it comes to the PIGs? Specifically, why have American firms been so willing to sell default insurance to the Europeans, though they have not bought much PIG debt? And conversely, why have the Europeans systematically been so eager to buy insurance for their PIG debt, even at the very high price such insurance now commands? In essence, European firms have been betting that a PIG default will happen sooner rather than later, while US firms have been betting that default would happen later or not at all.
There may be some subtle institutional reason for the dramatic difference in behavior between US and European financial institutions; I would welcome any insights or suggestions. An alternative explanation could be that there are informational differences, and/or that European financial institutions have a systematically different view of the PIGs than US financial institutions do. Could the Europeans know something that the Americans don't about the likelihood or timing of eventual default? I hesitate to believe that, but I don't have another good explanation for the one-sidedness of the betting on a PIG default...
UPDATE (June 13): In response to questions from a number of readers, here are some details about BIS definitions.
First note that these exposures are net, not gross exposures. Here's the BIS definition of exposures on a consolidated basis as reported in this data: "The consolidated banking statistics report banks’ on-balance sheet financial claims (ie contractual lending) vis-à-vis the rest of the world and provide a measure of the risk exposures of lenders’ national banking systems. The data cover contractual (immediate borrower) and ultimate risk lending by the head office and all its branches and subsidiaries on a worldwide consolidated basis, net of inter-office accounts... [T]o reflect the fact that banks’ country risk exposure can differ substantially from that of contractual lending due to the use of risk mitigants such as guarantees and collateral, reporting countries provide information on claims on an ultimate risk basis (i.e. contractual claims net of guarantees and collateral) since June 1999." (BIS Quarterly Review, Statistical Annex (pdf), p. A4.)
Second, there's some confusion about whether what I've labeled "indirect exposures" are really comprised significantly of credit default swaps (CDS). The confusion arises from the fact that the BIS data contains a line entry for "credit derivatives" which is very small for US banks (a net exposure of about $1bn for US banks with respect to Greece in that category), but then also has an entry call "guarantees" that is much larger (about $33bn for US banks with respect to Greece). It turns out that most CDS contracts are actually probably included in the "guarantees" line of the BIS data, not the "credit derivatives" line.
Here's the explanation from the BIS guide to their international statistics (pdf, p.14):
[C]redit derivatives, such as credit default swaps and total return swaps, are only reported under the item 'Derivative contracts' if they are held for trading by a protection-buying reporting bank. Credit derivatives that are not held for trading are reported as 'Risk transfers' by the protection buyer and all credit derivatives should be reported as 'Guarantees' by the protection seller.While we can't say for certain how much of US banks' indirect exposures to Greece are the result of CDS protection written on Greek bonds, we can say that any time a bank uses a CDS to take an open position rather than strictly for trading purposes (in which case we wouldn't expect the bank to have a significant open position by the end of the year anyway, which explains why the BIS entry for "credit derivatives" is so small), then that amount will show up in the BIS line called "guarantees". Given my suspicion that US banks are not writing a lot of letters of credit or otherwise directly issuing warranties and endoresements to Greek institutions, it seems likely that the bulk of the indirect exposure reported by the BIS for US banks to Greece is in fact in the form of CDS contracts.
'Guarantees' are contingent liabilities arising from an irrevocable obligation to pay a third-party beneficiary when a client fails to perform some contractual obligation. They include secured, bid and performance bonds, warranties and indemnities, confirmed documentary credits, irrevocable and standby letters of credit, acceptances and endorsements. Guarantees also include the contingent liabilities of the protection seller of credit derivative contracts.
UPDATE #2 (June 13): For more on this, see the new post Indirect US Exposure to the Euro Debt Crisis, part 2.