The concept of portfolio losses is fundamental to any quantification of credit risk. In order to properly account for counterparty risk at the portfolio level, there must be some statistical estimate of the possibility of significant losses in the event that many counterparties were to default in a given period of time (such as 1 year).

The quantification of counterparty risk at the portfolio level requires knowledge of the following factors:

- counterparty default probabilities;
- correlations between counterparty default events;
- randomness of future counterparty exposures;
- correlation between future counterparty exposures.

Whilst the first two components above are standard inputs for most credit portfolio models, the last two are specific to counterparty risk and create significant complications when treating OTC derivatives within a typical credit portfolio framework. Whilst the randomness and correlations of exposures can be assessed accurately with detailed knowledge of the relevant transactions with each counterparty, it may also be important to be able to treat these components in a simple fashion to avoid complex calculations.

We start with a simple example, considering losses arising from default events of two counterparties A and B^{4} with a direct exposure of 100 to each and default probabilities of 10%. In Figure 11.6 we show the loss probabilities for the case of no correlation and a high correlation value of ...

Start Free Trial

No credit card required