2014年-IMF国际货币组织全球_Market_Signals_and_the_Cost_of_Credit_Risk_Protection_An_Analysis_of_CDS_Settlement_Auctions_32页_589kb

Summary of "Market Signals and the Cost of Credit Risk Protection: An Analysis of CDS Settlement Auctions"

Core Content

This working paper investigates the relationship between Credit Default Swap (CDS) spreads and the actual recovery rates of defaulted bonds observed at CDS settlement auctions. The authors aim to assess whether the CDS spreads, which are used to infer default probabilities, accurately reflect the true probability of default based on post-default recovery data.

Main Points

• CDS as a Risk Indicator: CDS spreads are widely used to signal market expectations of default risk. They are often considered as a proxy for default probabilities, independent of other market factors like liquidity or microstructure.

• Settlement Auction Mechanism: CDS settlement auctions are used to determine the recovery value of defaulted reference obligations. These auctions are conducted by ISDA and managed by Creditex, and they are critical for resolving CDS contracts without physical delivery.

• Market Surprises: The authors observe that CDS spreads often do not align with the actual recovery rates seen at auctions. This implies that market participants may have been "surprised" by the actual default event, with lower CDS spreads corresponding to lower expected default probabilities than what was observed in reality.

• Excess Spreads: The paper introduces the concept of "excess spreads" as the difference between the CDS cost and the cost implied by the actual recovery rates. These excess spreads are influenced by the demand and supply mismatch in the auction process.

• Strategic Bidding Behavior: The authors suggest that auction participants may engage in strategic bidding, which can affect the final recovery prices and thus the implied default probabilities.

• Pricing Models and Recovery Assumptions: The paper highlights that CDS pricing models typically assume a constant recovery rate (e.g., 0.4 for senior unsecured bonds), which may not reflect the actual recovery rates observed at the auctions. This assumption can lead to mispricing of CDS contracts.

• Empirical Methodology: The authors use a dataset from CDS settlement auctions between 2005 and 2014, including final recovery prices, physical settlement requests, and net open interest (NOI). They compare implied default probabilities derived from CDS spreads with those derived from actual recovery values.

• Key Findings:

  • Implied default probabilities from CDS spreads are systematically different from those derived from actual recovery rates.
  • Excess spreads are significantly associated with market microstructure and macroeconomic conditions.
  • The use of standard recovery assumptions in CDS pricing may lead to mispricing of credit risk protection.
  • Auctions tend to be "one-sided," with high NOI indicating significant demand or supply imbalances.
  • The recovery rates for restructurings are higher than those for outright defaults, and the data supports the notion that recovery rates vary with the type of credit event.

Key Information

• Sample Size: 116 reference entities with credit events between 2005 and 2014.
• Auction Types: 73 cases of bankruptcy, 35 cases of failure to pay, and 32 cases of restructuring.
• Recovery Rates:
  • Senior bonds: average 42.74%
  • First lien loans: average 51.47%
  • Restructurings: higher recovery rates compared to bankruptcy and failure to pay.
• NOI Observations:
  • 93 out of 156 auctions had a negative NOI, indicating a strong interest in selling.
  • NOI is highly correlated with physical settlement volume (0.99), suggesting one-sided auctions.
• Recovery Assumption:
  • Standard recovery rate used in CDS pricing is 0.4 (for senior unsecured bonds).
  • This assumption is challenged by the authors, as actual recovery rates from auctions differ significantly.

Implications

• CDS Pricing Models: The assumption of constant recovery rates may not be appropriate, as it leads to discrepancies in implied default probabilities.
• Market Signals: The CDS spreads may not accurately reflect the true risk, as they are influenced by auction dynamics and strategic behavior.
• Policy and Practice: The findings suggest that policymakers and market participants should be aware of the limitations of CDS spreads in capturing true default risk, especially in the context of settlement auctions.

Conclusion

The paper concludes that the CDS market's pricing of credit risk protection may not be accurate due to the discrepancy between the assumed recovery rates and the actual recovery values observed at auctions. This has important implications for the reliability of CDS spreads as signals of default risk and for the pricing models used in the CDS market.