Credit risk
What is probability of default (PD)?
Probability of default (PD) is the estimated likelihood that a borrower will default on its obligations within a specified time horizon, usually one year.
Formula
Expected loss = PD x LGD x EAD
Probability of default estimates how likely a borrower is to default over a defined period, most often twelve months. Default is defined by the institution, commonly as 90 days past due, nonaccrual status, charge-off or bankruptcy. PD is a property of the borrower, and it is one of the three components of expected loss alongside loss given default and exposure at default.
Lenders estimate PD through statistical scorecards, internal rating models mapped to historical default rates, or external data. Each internal risk grade is often associated with a PD so ratings can feed loss estimates, pricing and capital calculations. PD estimates may be point-in-time, reflecting current conditions, or through-the-cycle, reflecting average conditions across an economic cycle.
PD models need enough default history to be reliable, which is a challenge for low-default portfolios such as large commercial loans. Inconsistent default definitions between model development and use, models that are not recalibrated as conditions change, and confusing point-in-time and through-the-cycle estimates can all produce misleading results. Models should be validated and back-tested against actual default rates.
See it in Risk OS
Loan risk rating software that keeps the reviewer's grade →
