But it's a bit frustrating not to have a clearer understanding of exactly what form this exposure takes. So I've been trying to see if there is any public information that can give us a hint about exactly how the big US banks have incurred such exposure.
Unfortunately, it's very difficult to get any good information about banks' derivatives exposures. The major US banks tend to downplay their exposure to the Euro debt crisis in their SEC filings. For example, in their 2010 10-K filing Bank of America wrote:
Certain European countries, including Greece, Ireland, Italy, Portugal and Spain, are currently experiencing varying degrees of financial stress. These countries have had certain credit ratings lowered by ratings services during 2010. Risks from the debt crisis in Europe could result in a disruption of the financial markets which could have a detrimental impact on the global economic recovery and sovereign and non-sovereign debt in these countries. The table below shows our direct sovereign and non-sovereign exposures, excluding consumer credit card exposure, in these countries at December 31, 2010. The total exposure to these countries was $15.8 billion at December 31, 2010 compared to $25.5 billion at December 31, 2009.In fact, B of A's direct exposure to Greece is listed at only about $500 million. But note that that is their direct exposure. What we learned from the BIS data is that they also have indirect exposures, which probably arise primarily through their credit derivatives purchases and sales.
The following table summarizes all of the useful information I could gather about credit default derivative activities from the 2010 annual SEC filings of the six largest banks in the US. It's not much, and doesn't really answer our original question, so mainly this just confirms what we have no way of knowing which US banks are exposed to the Euro debt crisis or why.
Let me walk you through the figures in this table. The first line shows CDS contracts sold by each bank -- essentially how much insurance against default each bank sold to third parties. Note that some of that insurance was on non-investment grade assets (which would include Greece, of course). That's the second line.
Now, it's often the case that a bank doesn't want to have an open position on that contract -- in other words, it doesn't really want to make the bet that the underlying asset will remain sound. In those cases the banks purchase default insurance themselves on the same asset. The effect is that the bank is then just making money on the spreads and fees, but has nothing to gain or lose based on whether the underlying asset defaults.
If you subtract line 3 from line 1, then you have the net open positions taken by each bank on the credit default insurance it has written. That line is highlighted in the table above, because that tells us how active each bank has been in actually placing bets on default outcomes. No details are provided by the banks about where those exposures are, however, either geographically or by type of underlying asset.
Finally, the last line is also somewhat interesting, because it tells us how much income each bank earned from their trading in CDS as well as their betting (taking open positions) on default events. Obviously, if you do it right you can really make some good money in the CDS market -- over $9 billion in 2010 for B of A. Note that every month that goes by without a Greek default is great for the sellers of default insurance on Greek assets, so it could well be that some of these banks are making a killing right now on the default insurance they've sold.
What can we conclude from this? I can think of a few things.
1. Bank of America, Morgan Stanley, and Goldman Sachs are the most aggressive in terms of taking open positions on default outcomes. But we have absolutely no idea how much of those positions (if any) were with peripheral Euro assets. Also, while the last two firms don't break out income attributable to CDS activities (at least not that I could find), B of A made a huge portion of their profits in 2010 from them. (Note that Citi did not indicate how much of the CDS protection that they sold was covered by purchases of CDS insurance, so they may or may not be in that list as well.)
2. The aggregate CDS exposures of the big US banks are certainly large enough to be plausibly consistent with the BIS estimate of about $100 bn in indirect exposures to peripheral Europe. If you add up the highlighted numbers (and make a guess at Citi's position), it seems reasonable to guess that the total net open positions on CDS protection sold to third parties by the big US banks is between $1,500 and $2,000billion. Attributing $35 bn of that (about 2%) to Greece, which has certainly had one of the most active markets (proportionally) for CDS contracts over the past year, doesn't seem to be a stretch.
3. Banks do not have to provide much detail about the indirect credit exposures that they take on when they sell default insurance through the CDS market. We have incredibly scant information about the positions that US banks take through default insurance, and therefore no idea about how any individual bank will be affected by a Greek default.
4. It's hard to find any other potential exposures to Greece, Ireland, and Portugal in the banks' public filings, other than through CDS contracts. Combined with points 2 and 3 above, the process of elimination suggests to me that CDS contracts are indeed likely to be the source of the bulk of US banks' indirect exposures to a Euro-zone default.
Finally, a note about the risk this poses to the US banking system. The big US banks are well-capitalized now, and can fairly easily absorb losses of several billions of dollars in the event of a Greek default. But two serious concerns remain. First, I fear that this may have the potential consequence of exacerbating the flight to safety that will happen in the event of Greece's default; if you have no idea who is really going to be on the hook and ultimately liable for CDS payments, your best strategy may be to trust no one. I don't think that triggering post-traumatic flashbacks of the fall of 2008 is going to do good things to the market or the economy. Second, I wonder if there's a public relations disaster just lying in wait for the big US banks. After all, how will you feel (assuming you don't work on Wall Street) when you read the headline that Big Bank X lost money because it sold billions of dollars of credit default insurance while it was on taxpayer life-support? Rightly or wrongly, I'm guessing that Big Bank X will not be very popular for a while.