Explain PD, LGD, and EAD, and how they combine into expected loss for credit risk.

These are the building blocks of credit-loss measurement. PD (probability of default) is the likelihood the borrower defaults over a horizon (typically one year for regulatory, or lifetime for IFRS 9 stages). LGD (loss given default) is the proportion of exposure you actually lose after recoveries and collateral — i.e., 1 minus the recovery rate. EAD (exposure at default) is the amount outstanding expected to be owed at the time of default, which for undrawn lines includes a credit-conversion factor for likely drawdown before default. Expected loss = PD × LGD × EAD. The framework matters because each parameter is estimated and validated separately, they drive both regulatory capital (IRB) and IFRS 9 ECL provisioning, and they're sensitive to the cycle — PD rises and LGD worsens (collateral values fall) together in a downturn, which is the correlation that makes credit losses spike. Unexpected loss (the tail beyond EL) is what capital covers.