Inconveniently they have chosen 15 highly liquid financial names for comparison purposes, which unfortunately means their conclusions do not address our concerns that credit default swaps (CDS) spreads remain questionable estimates when the CDS and the rated underlying itself have vastly different trading volumes, or for illiquid or unrated securities (visit our April piece "Credit Ratings vs. Credit Default Swaps").
As the banking regulators consider following the NAIC's lead in finding alternative solutions to relying on credit rating agencies for regulatory capital reserve considerations, another key features to consider is their respective predictive content: do ratings or CDS spreads have any long term opinion associated with them, or are they purely back-looking or point-in-time estimates. The authors attend directly to the search for regulatory scrutiny alternative and the possibility of relying on CDS for this purpose (emphasis added by us):
"More generally, it is apparent that CDS spreads reflect available information, which makes them useful for regulatory and risk management purposes, even if they are not necessarily suitable for forecasting."
"At a minimum, our analysis supports the conclusion that CDS spreads reflect information more quickly and accurately than credit ratings. Specifically, we find that as information about the subprime mortgage exposure of financial institutions was disclosed during 2007 and 2008, CDS spreads reflected that information, whereas credit ratings remained relatively unchanged.
If regulators and investors had looked to CDS spreads to assess the riskiness of financial institutions during this period, they would have found as early as April 2007 that such risks were significant and increasing. By early 2008, CDS spreads reflected a significant likelihood of default by one or more investment banks. In contrast, credit ratings reflected little or none of this information."