Credit risk is the possibility that a borrower will default on a loan. In this context, default is defined broadly as any time the borrower does not meet the terms of his contractual obligations with the lender. Because lenders are inclined to provide credit only to those who are likely to repay them, borrowers usually do meet their obligations. Thus, obligor defaults are relatively rare, and credit risk analysis is made difficult by the scarcity of historical data available to use for computing the probability of default (PD) and loss given default (LGD).
Further complications arise because of dependence among obligors, whose abilities to repay their debts are affected to some degree by other borrowers doing business in the same economy. This contagion among borrowers manifests itself in statistical correlations between defaults, which must be incorporated into credit risk models. Finally, PDs and LGDs can change over time due to changes in general economic factors and other conditions affecting individual borrowers.
This chapter describes a Crystal Ball model that is used to simulate correlated defaults, and produce the distribution of potential losses for credit risk analysis and management. For more information about credit risk modeling, see Bluhm, et al. (2003), Crouhy, et al. (2001), de Servigny and Renault (2004), or Ganguin and Bilardello (2005).
14.1 EXPECTED LOSS
The first thing to know about credit risk analysis is how to compute expected credit ...