Credit loss distribution: Monte Carlo
About this app
Simulate a credit portfolio’s loss distribution and explore expected loss, VaR, CVaR and the effect of default correlation on the tail.
The idea
The loss on a credit portfolio is random: each borrower defaults with probability PD, and on default the fraction LGD of the exposure EAD is lost. The expected loss EL = Σ EAD · PD · LGD is the mean of the loss distribution. Value at Risk (VaR) is its quantile at a confidence level such as 99 %, and Conditional VaR (CVaR) is the average loss beyond the VaR. The gap VaR − EL is called economic capital, the buffer a bank holds against unexpected losses.
What you see
The app simulates a one-factor model: each borrower’s asset return is Aᵢ = √ρ · M + √(1 − ρ) · Zᵢ with a shared factor M, and the borrower defaults when Aᵢ falls below the threshold Φ⁻¹(PD). Every scenario adds one portfolio loss to the histogram, and markers show EL, VaR, CVaR and economic capital. In “dice” mode each borrower defaults with probability 1/6 and the app simply counts defaults.
Try it
Set the number of borrowers (5 to 200), PD, LGD and the confidence level (95 %, 99 % or 99.9 %), then press “Auto”. Now raise the correlation ρ: the expected loss stays the same, but the right tail gets heavier and VaR and CVaR move to the right. Switch correlation off to see how narrow the distribution stays when defaults are independent.
Frequently asked questions
- What is the difference between expected and unexpected loss in credit risk?
- Expected loss (EL) is the average loss a bank anticipates over time and builds into its loan pricing. Unexpected loss describes how far a bad year can exceed that average and is usually measured as VaR − EL. Banks hold capital to absorb it.
- How does CVaR differ from VaR?
- Value at Risk is the loss that will not be exceeded with a chosen probability, for example 99 %. Conditional VaR, also called expected shortfall, is the average loss in the cases where the VaR is exceeded. It therefore also reflects how severe the extreme scenarios are.
- Why does default correlation make a loan portfolio riskier?
- When defaults are correlated, many borrowers fail together in bad scenarios. The mean of the loss distribution does not change, but its right tail becomes heavier. VaR, CVaR and the economic capital required all go up.
- What can I adjust in the loss distribution app?
- You choose portfolio or dice mode and set the number of borrowers, PD, LGD, the correlation ρ and the VaR confidence level. The simulation runs step by step or automatically, with EL, VaR, CVaR and economic capital updated as it goes.
Subject: Credit risk
More from Credit risk
- Adverse selection
- Default correlation surface in 3D
- CreditPortfolioView: macro factors
- CreditRisk+: Poisson modelling
- Exposure profile
- Integrated risk measurement
- Rating migration explorer
- PD master scale
A VisuApp by heyprof: interactive, free in your browser, no sign-up. What is a VisuApp? · All apps