Expert judgment boosts Banca d'Italia's credit risk assessment accuracy
A Banca d'Italia study reveals that expert assessment significantly improves the predictive power of its in-house credit assessment system (ICAS) for Italian non-financial firms. The methodology now integrates climate-related risks and sector analysis, increasing AUROC by 2 percentage points.
Beyond algorithms: The human touch in credit ratings
Banca d'Italia's In-house Credit Assessment System (ICAS) plays a crucial role in monetary policy implementation by ensuring the quality of credit claims pledged as collateral.
ICAS comprises a statistical model and an expert assessment, applied to significant firms for collateral purposes.
The expert assessment, conducted by two credit analysts, introduces a forward-looking perspective and qualitative judgment, addressing the limitations of purely statistical approaches.
Analysts independently review firm characteristics and assess predefined risk profiles, or 'assessment modules', which are then weighted to determine the final 'full rating'.
Recent methodological enhancements include the integration of climate-related risks and major changes in sector analysis.
This comprehensive review, carried out in 2023, aims to increase the homogeneity, reliability, and transparency of the qualitative credit assessment.
Empirical proof: Enhancing prediction and resilience
Leveraging over 25,000 assessments from 2016 to 2022, the study empirically validates the expert assessment's contribution.
It significantly enhances both the predictive and discriminatory power of the full ratings, improving the AUROC by approximately 2 percentage points.
This gain is crucial, especially as the expert assessment protects ICAS performance during periods of macroeconomic stress.
The methodology has undergone continuous improvements, including a comprehensive review in 2023, to ensure greater homogeneity, reliability, and transparency in qualitative credit assessment.
This ongoing refinement ensures the system's robustness against heightened uncertainty.
Human judgment: An indispensable safeguard
Human expert judgment is indispensable for robust credit risk models, offering a crucial layer of resilience.
Its ability to integrate forward-looking qualitative factors and emerging risks, like climate change, provides an essential safeguard against purely quantitative limitations.
For central banks, this validates the enduring value of skilled analysts in maintaining financial stability and effective monetary policy transmission.