Quiet Books, Compressing Margins: An 18-Bank Cohort Carrying Rate Risk Below the Screens

The cohort numbers 18 institutions, drawn from a universe of 4,108 active banks (0.44%). Asset-band distribution skews small: 8 institutions under $200M, 3 in the $200M-$500M band, 6 in the $500M-$2B band, and 1 in the $10B-$50B band. Geographic concentration is pronounced. New Mexico (4), Oklahoma (3), Texas (3), Missouri (2), and Wisconsin (1) together account for 72.2% of cohort membership — a Southern Plains and Mountain West footprint rather than a national scatter.

HTM exposure across the cohort spans a median of 29.17% of assets, with a 25th-75th percentile range of 25.35% to 41.44% and a maximum of 67.85%. The upper tail is where the profile becomes vivid. First State Bank, New Mexico (FDIC CERT 16600) carries HTM at 67.85% of a $208.7M balance sheet, with core deposits at 87.4% and NIM down 64 basis points year-over-year. Community State Bank, Oklahoma (CERT 27201) shows HTM at 57.33% against a $79.5M book, core deposits at 91.6%. Park Bank, National Association, Wisconsin (CERT 13054) carries HTM at 52.95% of $79.7M in assets. The Falls City National Bank, Texas (CERT 3193), at $591.3M, holds HTM at 50.37% with NIM compression of 29 basis points. None of these institutions would flag on a conventional CRE concentration, asset quality, or capital adequacy screen. Their surface metrics read as conservative community banking.

The v0.1 precedent corpus does not contain a direct analog for this cohort, and honest treatment requires saying so. The structural pattern — long-duration securities books funded by sticky deposits, compressed on the margin by a rate cycle — is the shape that stressed several regional banks in 2022-2023, but those episodes involved institutions of materially different scale, business mix, and uninsured deposit posture than the community banks that populate this cohort. Extrapolating from those events to a Falls City or Okarche would misrepresent the mechanism. Pattern matching at this corpus depth is illustrative, not statistically rigorous. The mechanism analysis below carries the analytical weight for this cohort.

The structural exposure works as follows. Each of the four defining metrics — core deposit ratio above 80%, fee income contribution below 10%, HTM above 20% of assets, and year-over-year NIM decline — is individually benign or actively favorable under standard supervisory frameworks. High core deposits imply low reliance on brokered or wholesale funding, an unambiguous strength in liquidity assessment. Low fee income indicates a traditional spread-lending model, which examiners generally read as low complexity rather than fragility. HTM classification under ASC 320 permits amortized-cost accounting, so unrealized losses on rate-driven securities repricing do not flow through regulatory capital for non-AOCI-opt-in institutions. NIM compression is common across community banks in the current cycle and is not, by itself, a supervisory flag.

The compound is where the exposure lives. A bank that funds a large HTM book with core deposits has, in substance, term-transformed short-duration liabilities into long-duration assets whose economic value has repriced but whose carrying value has not. The core deposit franchise is the ballast — if depositors remain sticky, the HTM portfolio pulls to par over time. If margin compression persists, the earnings capacity to absorb any credit or operational stress narrows. If deposit behavior shifts, the option to hold securities to maturity depends on liquidity that may not be there. The 2006 Interagency Guidance on Concentrations does not speak to securities concentration in the way it speaks to CRE, and there is no equivalent screen. That gap is precisely why the profile does not surface.

CAMELS engagement here is concentrated in L (liquidity) and S (sensitivity to market risk). The relevant supervisory dialogue is about interest rate risk modeling, deposit beta assumptions, and the realism of hold-to-maturity intent under stress — not asset quality or capital ratios read at par.

Klaros should treat this cohort as a Discovery Brief candidate with three action arcs.

1. Engagement positioning. The 18 named institutions, and particularly the upper-tail members (First State Bank NM, Community State Bank OK, Park Bank NA WI, Falls City National TX), are candidates for outreach framed around interest rate risk governance and HTM portfolio strategy rather than around distress. The value proposition to management is anticipatory: examiners are increasingly attentive to the compound profile even where individual metrics do not trigger, and boards benefit from a structured IRR narrative before that conversation arrives.

2. Examination preparation advisory. For cohort members already in an examination cycle, Klaros can help management articulate deposit beta assumptions, HTM intent documentation, and NIM trajectory in terms that anticipate S-component and L-component questions. The regulatory framework does not screen for this profile, but individual examiners are increasingly asking about it, and preparation is asymmetric in value.

3. Cohort tracking. The Sector Pulse Engine should re-run this cohort each quarter to test persistence. If the 18-member roster is stable across 2026Q1 and 2026Q2, the Discovery Brief becomes publishable — a named cohort that CharterIQ surfaced before conventional analysis. If the roster churns significantly, the profile is transitory and the mechanism claim weakens. Persistence is the test that converts a candidate cohort into doctrine.