Credit Risk Summit
   

New York, April 9th 2010                 

CONFERENCE PROGRAM 

Pierre COLLIN-DUFRESNE, Columbia University

On the Relative Pricing of long Maturity S&P 500 Index Options and CDX Tranches

We investigate a structural model of market and firm-level dynamics in order to jointly price long-dated S\&P 500 options and tranche spreads on the five-year CDX index.  We demonstrate the importance of calibrating the model to match the entire term structure of CDX index spreads because it contains pertinent information regarding the timing of expected defaults and the specification of idiosyncratic dynamics. Our model matches the time series of tranche spreads well, both before and during the financial crisis, thus offering a resolution to the puzzle reported by Coval, Jurek and Stafford (2009a).

Rama CONT,   Columbia University

Are CDO models any good for hedging? An empirical study
Joint work with: Yu Hang KAN  (Columbia University)
We compare the performance of various hedging  strategies for index CDO tranches across a variety of models and hedging methods before and during the recent credit crisis. Our  analysis shows poor performance of most model-based delta-and gamma-hedging strategies and strong evidence for unhedgeable risk in CDO tranches. We also show that, unlike what is commonly assumed, dynamic models do not necessarily perform better the static models, nor do high-dimensional default correlation models perform better than simpler top-down models. In fact hedges based on top-down models and regression-based betas outperform significantly sensitivity-based hedging strategies based on single name CDS during the Lehman Brothers default event.

Saiyid ISLAM and William MOROKOFF, Standard & Poor's  

Measuring Performance of Credit Risk Assessments

Popular measures for evaluating the performance of various credit risk measurement models have traditionally relied on computation of Lorenz curve-based scores such as Accuracy Ratios or Gini Coefficients. These measures are also commonly used to assess default risk predictions across very different sectors, regions or time periods. Such measures are however sensitive to sample size and to realized default rates that may vary substantially across economic cycles or sectors due to correlation effects. We analyze the performance of these performance measures themselves through a controlled study of real and modeled data and present results related to the impact of small data sets, datasets with few or zero defaults, and datasets where the
obligations are highly correlated to gauge the effectiveness of these traditional credit risk performance measures. We also highlight instances where the performance of traditional measures may be difficult
to interpret.

Dilip MADAN, Robert H. Smith School of Business

Accounting to Acceptability: With applications to the pricing of ones own credit risk.

The theory of pricing to acceptability developed for incomplete markets by Cherny and Madan (2009b) is applied to marking ones own default risk. It is observed in agreement with Heckman (2004), that assets and liabilities are not to be priced under fair value accounting principles at the same magnitude. Liabilities are marked at ask prices that are above the asset mark at bid prices.  Applying cones of acceptability defined by the concave distortion minmaxvar at the stress level of 0.75 it is shown that counterintuitive profitability resulting from credit deterioration is eliminated. Following Heckman we suggest that the difference between the liability mark and the asset mark be taken as an upfront expense deposited in a special account called the ODOR account for Own Default Operating Reserve. Procedures for pricing coupon bonds separately as assets and liabilities are described. They employ the default time distribution as calibrated from the CDS market.

Richard MARTIN, Man Investments

The Merton model revisited: Use of local Levy processes for understanding the credit-equity relationship

The Merton, or structural, model has been extended in different ways in recent years, commonly through the use of a barrier option to represent
the optionality present in the firm's debt or CDS. It has been known for some time that a simple diffusion model for the firm's assets is
incapable of fitting the CDS term structure seen in the market, and that some kind of jump process is required. What is less well known is that a
local volatility needs to be introduced to capture the credit-equity relationship properly. I will explain why this is and how to do the modelling and computation.

Cocktail Reception - Room 1