Credit Risk Modeling: PD, LGD, EAD & Expected Credit Loss

What you'll build

What you will be able to do

Each lab ends with a concrete result that you build yourself and the checker confirms, cell by cell.

• Calculate the expected loss of every loan in Harrow Valley Bank's sample book, the exposure-weighted portfolio loss rate and a stressed expected loss.
• Estimate a credit conversion factor from five past defaults, then use it to calculate the exposure at default of Harrow Valley Bank's current revolving credit lines.
• Estimate the loss given default of eight Harrow Valley Bank loans from their collateral, haircuts, workout costs and the time value of the recovery.
• Roll a B-rated borrower through Harrow Valley Bank's one-year rating migration matrix for four years and derive the cumulative, marginal and average annual PD.
• Build the expected credit loss of a five-year amortising loan year by year, compare the 12-month and the lifetime ECL and weight three economic scenarios.
• Assign ten Harrow Valley Bank loans to IFRS 9 stages with a PD test, the 30-days-past-due backstop and a watch-list flag, then book the right ECL per stage.
• Score eight loan applicants with a logistic regression model, convert the log-odds into PDs and into scorecard points with PDO scaling, and count approvals at a cut-off.
• Summarise Harrow Valley Bank's loan book by segment: gross exposure, ECL allowance, coverage ratio and the share in stages 2 and 3, and chart the coverage.