What is the bond/CDS basis?
Accrued Interest has a fascinating data point on the troubles of the financials at the moment: Citi's latest five year was done at 337 over Treasuries or roughly 235 over Libor.
AI then suggests that there is an arb between CDS at 160 over and the bond at 239 over. In ordinary conditions that would be true because in ordinary conditions a trader would assume he or she could fund the bond at Libor flat, buy the CDS at 160, be risk flat, and make 79 running. But... firstly at the moment there aren't many Libor flat funders. The C deal discussed is evidence enough of that. And second you are only risk free if the CDS is guaranteed to pay - and which counterparty can be sure of that for in a situation where Citigroup is in default? The 160 over then does not reflect the real cost of being sure of protecting Citi debt, but rather the credit weighted average cost across typical CDS counterparties: buying protection on C from Berkshire Hathaway, say, would probably be more expensive (it certainly ought to be if Warren's folks are sharp).
A trade like long the debt long CDS protection only makes sense if (Bond spread over Libor - my funding cost over Libor to term) > (CDS spread + cost of complete credit protection on CDS counterparty for CDS receivable). The 2nd and 4th terms are typically small in an ordinary market. But this ain't no ordinary market.
AI then suggests that there is an arb between CDS at 160 over and the bond at 239 over. In ordinary conditions that would be true because in ordinary conditions a trader would assume he or she could fund the bond at Libor flat, buy the CDS at 160, be risk flat, and make 79 running. But... firstly at the moment there aren't many Libor flat funders. The C deal discussed is evidence enough of that. And second you are only risk free if the CDS is guaranteed to pay - and which counterparty can be sure of that for in a situation where Citigroup is in default? The 160 over then does not reflect the real cost of being sure of protecting Citi debt, but rather the credit weighted average cost across typical CDS counterparties: buying protection on C from Berkshire Hathaway, say, would probably be more expensive (it certainly ought to be if Warren's folks are sharp).
A trade like long the debt long CDS protection only makes sense if (Bond spread over Libor - my funding cost over Libor to term) > (CDS spread + cost of complete credit protection on CDS counterparty for CDS receivable). The 2nd and 4th terms are typically small in an ordinary market. But this ain't no ordinary market.
Labels: CDS, Liquidity risk
posted by David Murphy at 11:31
0 Comments:
Post a Comment
<< Home