Credit Scoring & Probability of Default Flashcards
7 cards from real CCP practice questions. Tap to flip, then mark Knew It or Still Learning — missed cards come back until you master them.
Read the first 7 Credit Scoring & Probability of Default flashcards as text
What is 'adverse action' as it relates to credit scoring under the Equal Credit Opportunity Act (ECOA)?
Answer: A denial, termination, or unfavorable change in credit terms, requiring reason codes to be provided to the applicant
ECOA requires lenders to provide specific reason codes (adverse action reasons) explaining why credit was denied or terms were changed unfavorably.
A Merton-style structural credit model estimates PD by treating the firm's equity as a call option. What is the 'default boundary' in this framework?
Answer: The asset value level below which the firm cannot service its debt obligations
In the Merton model, default occurs when the firm's asset value falls below the value of its liabilities (the default boundary) at debt maturity.
Which of the following scorecard development practices violates fair lending principles in the United States?
Answer: Using race or national origin as a scorecard characteristic
ECOA and the Fair Housing Act prohibit using race, color, national origin, sex, religion, or other protected characteristics in credit scoring or underwriting decisions.
A credit risk manager is reviewing a model's calibration. The model predicts an average PD of 2.5% for a risk band, but the observed default rate is 4.1%. What does this imply?
Answer: The model is underpredicting default risk, meaning capital reserves may be understated
When observed defaults exceed model-predicted PDs, the model is underpredicting risk, which could lead to inadequate capital reserves and underpriced credit.
In retail credit scoring, what is the 'bad definition' and why is it critical?
Answer: The performance outcome (e.g., 90+ days past due) used to label accounts as 'bad' during model development
The bad definition specifies the delinquency threshold (commonly 90+ DPD within 24 months) used to classify accounts as defaulted in the training dataset, directly shaping the model's target variable.
A bank uses a cut-off score of 650 for auto loan approvals. Which of the following best describes the trade-off when lowering the cut-off to 620?
Answer: Approval volume increases but expected losses also increase
Lowering the cut-off score approves riskier applicants, increasing both approval volume and expected credit losses, requiring a risk/return trade-off analysis.
Which of the following best describes 'model risk' in the context of credit scoring?
Answer: The potential for adverse consequences arising from decisions based on an inaccurate or misused model
Model risk, as defined by the Fed's SR 11-7 guidance, is the risk of losses from reliance on incorrect, misapplied, or poorly validated models used in decision-making.