A CVA engine needs the value of every live trade in the same future state, the collateral already settled in that state, and the default event that turns a claim into a loss. Drawing future market prices is only the first part. This lesson assembles exposure profiles, netting, collateral and bilateral default into a reproducible simulation, then checks a small wrong-way-risk example against an exact answer.
Learning objectives. By the end of this note you should be able to:
- distinguish outer market scenarios from conditional trade revaluation;
- implement netting and collateral in the correct sequence;
- estimate CVA and its sampling error from pathwise losses;
- distinguish direct default simulation from conditional default weighting;
- diagnose sampling error, grid bias and valuation bias separately.
In this lesson
- Define the object before drawing paths
- Revaluation is a conditional pricing problem
- The computational sequence
- Direct simulation versus conditional weighting
- A runnable model with an exact answer
- Error accounting and reproducibility
Before you start
The full lesson, with worked examples and exercise solutions, is for members.
Read the full lessonNew to SelfMadeQuant.com? See membership options or browse the Counterparty Credit Risk and CVA syllabus.