Summary Points
-
Credit risk starts with uncertainty: banks lend without knowing who will repay, so they group similar loans and assess the risk of default.
-
Three parameters capture potential loss: PD (chance of default), LGD (share lost), and EAD (amount owed); together, they estimate expected loss.
-
Models turn past data into risk estimates: define default, build reliable datasets, rank borrowers, calibrate probabilities—and validate, document, and monitor the results.
-
Models support safer lending and capital: Basel rules govern how estimates are used to calculate risk-weighted assets and help banks withstand losses beyond the expected average.
Risk Is a Question, Not a Prediction
A bank lends today without knowing what a borrower’s finances will look like years from now. Credit risk measures the chance that a borrower will fail to meet the loan’s terms—and the loss the bank could face if that happens.
Analysts estimate three things. Probability of default (PD) measures the chance of default, often over one year. Loss given default (LGD) estimates how much of the debt the bank may not recover. Exposure at default (EAD) estimates how much the borrower will owe then.
Together, these measures estimate average losses. For example, a $250,000 loan with a 2% PD and 40% LGD has an estimated expected loss of $2,000, assuming the full loan remains outstanding. That does not mean the bank will lose exactly $2,000—or that the borrower will repay 98% of the loan. It is an estimate across comparable cases.
Data Gives the Model Its Meaning
To estimate risk, a bank studies past borrowers. It records what it knew about them at a given point, then checks whether they defaulted during the following year. Analysts use those patterns to group borrowers by risk and estimate probabilities.
That process demands care. Missing records can skew a sample. A calm economic stretch may hide risks that emerge in a downturn. And a model that ranks borrowers well may still assign probabilities poorly. So teams test results, document assumptions and monitor performance as conditions change.
The Basel framework sets prudential standards for how banks measure risk and hold capital. Some banks apply prescribed risk weights. Others, with supervisory approval, use internal estimates for certain risk measures. Internal models can reflect a bank’s portfolio more closely, but they also require strong data, controls and oversight.
Useful Measures, With Limits
Expected loss helps banks price loans and recognize estimated losses in their accounts. Capital serves a different purpose: It helps absorb losses beyond the average, including severe losses during economic stress. Regulators use risk measures to inform capital requirements and limit the chance that trouble at one bank spreads across the system.
Still, a model cannot foretell which borrower will default or when. Its value lies elsewhere. It gives banks a consistent way to compare risk, make decisions and prepare for losses—provided they keep testing its assumptions and limits.
Continue Your Tech Journey
Stay informed on the revolutionary breakthroughs in Quantum Computing research.
Stay inspired by the vast knowledge available on Wikipedia.
AITechV1
