Qualitative data and sector models anchor large firm credit ratings
BDI Paper

Qualitative data and sector models anchor large firm credit ratings

Credit rating frameworks for large corporations rely on qualitative expert judgement and sector models across European institutions, a Banca d'Italia study shows. The comparative review highlights common group constraints alongside divergent integration of artificial intelligence and ESG metrics.

Shared foundations across four rating pillars

The paper reviews credit assessment frameworks across four channels: rating agencies such as Moody’s, S&P, and Fitch, bank Internal Ratings-Based models, Italian agencies Cerved and CRIF, and central bank In-house Credit Assessment Systems.

Across all frameworks, hard balance-sheet metrics are systematically paired with soft information on governance, competitive positioning, and management interviews.

Methodological commonalities include sector-specific calibrations for capital-intensive industries like real estate and energy, formal group-level evaluations that cap subsidiary ratings at the parent level, and structured committee procedures for model overrides.

Stalled by black boxes and data gaps

Divergence appears in sustainability and technology.

While ICASes implement mandatory ECB climate standards, ESG factors across all providers remain qualitative overlays rather than statistical model inputs.

Artificial intelligence is similarly restricted to data screening and text mining due to explainability constraints and EU AI Act compliance.

Operationally, the 2026 Eurosystem credit framework reform moves SME assessment to automated models, focusing central bank expert analysis entirely on large corporations.

Human craft in an algorithmic age

Rating large firms remains a fundamentally artisanal exercise that algorithms cannot yet replace.

Yet relegating ESG risks and machine learning to qualitative sidecars exposes credit assessment systems to blind spots.

True comparability across Eurosystem collateral sources will remain out of reach until quantitative standards align.

Source: No. 94 - The Credit Assessment of Large Firms

IN:

Report an error