Explanation

Option A correctly describes an essential aspect of credit risk model development:

• Mapping credit scores to risk rating classes is a critical implementation step that ensures the model's outputs are actionable and meaningful for risk management decisions.
• Basing this mapping on empirical PD estimates ensures the model reflects actual default probabilities rather than arbitrary cutoffs.
• Ensuring well-diversified distribution across rating categories prevents concentration risk and allows for meaningful differentiation between risk levels.

Why other options are incorrect:

• Option B: Model validation involves much more than just external comparisons. It includes backtesting, sensitivity analysis, benchmarking against alternative models, and assessing model stability over time.

• Option C: Expert judgment is often essential in credit risk modeling, particularly for variable selection, model specification, and interpreting results. Pure algorithmic approaches may miss important business context.

• Option D: Credit risk models require continuous monitoring and periodic re-evaluation of borrowers. Ratings should be dynamic and updated as borrower circumstances change, not remain static.

The development process for credit risk scoring models typically involves multiple phases: data collection, variable selection, model development, validation, implementation, and ongoing monitoring - with option A correctly capturing a key implementation consideration.