Model methodsBankingactive

FinObservatory bank-health composite

A public-data, peer-relative bank-health composite on a 0–100 scale, where higher means financially stronger.

Registry slug:: bank-health-composite
Visible surface:: /banks

Data and implementation

Vintage

236,016 bank-quarter observations for 6,590 institutions, 2015Q1 through 2026Q1, measured 2026-08-30

Data tables

data/parquet/bank_scores.parquet

Engine

Producer

scripts/build_bank_scores.py

Outputs

data/parquet/bank_scores.parquet

Source: FDIC BankFind Suite, public call-report financials

Method

This is NOT a CAMELS rating.

The producer uses only public FDIC call-report fields. Within each quarter it assigns banks to one of five asset-size peer groups, orients every available raw metric so higher means stronger, converts each metric to an average-tie percentile within its peer-quarter, and averages the available metric percentiles inside Capital, Asset Quality, Earnings, Liquidity, and Sensitivity components.

The Sensitivity component is explicitly a growth-extremity proxy: unusually fast growth and unusually fast shrinkage both rank weakly relative to the peer-quarter median. Public summary fields do not support a duration gap or an unrealized-securities-loss-to-capital measure, so this component does not reproduce supervisory market-risk analysis.

The composite reweights across present components. It is published only when at least four of the five components are present; otherwise it is null rather than inferred from too little information.

Methodology evidence

Path and lines	Evidence
scripts/build_bank_scores.py:25-190	Defines the actual producer, public call-report inputs, peer groups, percentile construction, component weights, growth-extremity proxy, and minimum-component rule.
docs/banks_methodology.md:8-28	Defines the public-data proxy and the mandatory distinction from confidential supervisory CAMELS ratings.
docs/banks_methodology.md:70-186	Documents component metrics, peer-percentile orientation, the Sensitivity proxy, the printed composite formula, and missing-component treatment.

Equations

Composite

composite = (0.25*Capital + 0.25*AssetQuality + 0.20*Earnings + 0.15*Liquidity + 0.15*Sensitivity) / sum of weights present

Validation

The output contains 236,016 bank-quarter observations for 6,590 institutions from 2015Q1 through 2026Q1.
The producer writes the five components, the count of components present, and the composite beside the public filing identifiers, making the minimum-component rule auditable row by row.

Limitations

This public-data proxy cannot reproduce confidential examination evidence or supervisory CAMELS ratings.
Each score is relative to an asset-size peer group in the same quarter, not an absolute solvency threshold, and the largest peer group is coarse because it contains few banks.
The Sensitivity component measures asset-growth extremity, not interest-rate duration, securities marks, or deposit-run exposure. A high score is not a solvency guarantee and a low score is not a failure prediction.

References

FDIC BankFind Suite, public call-report financial data.

Metadata endpoint

Path	Method	Returns	Example
/api/methods/[slug]	GET	Registry metadata, implementation paths, measured vintage, methodology evidence, validation, and limitations. No model observations or parquet contents.	/api/methods/bank-health-composite