Peer Cohort (Size Band)

A peer cohort is the size band an institution is grouped into for comparative analysis — typically by total assets — so that a $400M community bank's performance is compared against other community banks rather than against trillion-dollar global institutions. Peer-cohort normalization is the foundation of any credible bank- or credit-union-distress ranking: a 6% Tier 1 leverage ratio means something very different at a $250M rural bank than at a $30B regional, and a flat ranking that lumps them together produces noise, not signal.

Why Compare Within a Peer Cohort?

Larger institutions typically run leaner capital ratios because they have richer access to capital markets and more diversified loan books. A 7% CET1 at a megabank is competitive; the same 7% at a community bank is borderline thin. Conversely, smaller institutions often carry concentrated geographic and industry exposures that justify thicker capital buffers but don't translate into headline-grade stress until a regional shock arrives.

Cohort-normalized ranking solves this by computing each institution's deviation from its cohort median:

Bank A and Bank B look identical on a flat ranking — both running 22% Texas Ratio. But Bank A is doing something materially worse than its peers, while Bank B's number reflects sector-wide pressure rather than institution-specific distress. The deviation metric isolates the institution-specific signal that is the relevant prediction for a note-investor pool-sale inquiry.

NPL Explorer Cohort Bands

FIXnotes' NPL Explorer uses four size bands for bank cohort normalization, calibrated to the natural distribution of FDIC-insured institutions:

The same four-band structure applies to credit unions sourced from NCUA Form 5300, though the population distribution is more right-skewed — most credit unions sit in the under-$100M band, and a $1B+ credit union is already a top-tier institution.

Cohort assignment is sticky within a quarter — institutions are bucketed once per call-report panel based on end-of-quarter total assets. Banks crossing a cohort boundary mid-trend (a $950M bank growing through $1B) move to the larger cohort at the next quarter-end.

Cohort Medians vs. Cohort Means

NPL Explorer ranks deviations against the cohort median, not the mean. Bank capital and distress metrics have heavily right-skewed distributions — a handful of severely distressed banks per quarter would pull the cohort mean significantly higher than the typical institution. The median is robust to these tail observations and represents the "typical bank in this cohort" more honestly.

Peer Cohort as Methodology

Cohort definitions are not regulator-issued conventions — the FDIC Bankers Resource Center does not publish an official size-band taxonomy for analytics, and the NCUA's regulation and supervision framework similarly leaves cohort selection to the analyst. Different analytics platforms use different bands — Federal Reserve uniform bank performance reports use one taxonomy, S&P Global uses another, FIXnotes uses the four bands above. The right choice depends on the analytical question; the only universal rule is that some size-band normalization is necessary for any cross-institution ranking to be interpretable.

For trend analysis paired with cohort comparison, see quarter-over-quarter (QoQ).

See current top 50 distressed banks normalized against peer cohort → Distressed vs Peer.