This is primarily for bank credit analysts writing a credit memo on U.S. banks with strong reported earnings and weak-looking credit quality. Fintech founders choosing a sponsor bank, financial journalists checking an earnings headline, and small-bank directors deciding what to ask management can use the same screen, but the main job is credit triage: when ROA is solid while nonperforming loans are high, should profitability be treated as earnings power or as a lagging indicator before credit costs arrive?
A clean ROA number can sit beside a messy credit file because Call Report earnings and asset-quality schedules move on different clocks. In this post, earnings means Schedule RI income and Schedule RC-K average assets; asset quality means Schedule RC-N past-due and nonaccrual loans, Schedule RC-C loan balances, Schedule RI-B charge-offs, Schedule RI-C allowance movement, and Schedule RC-R capital. After this point, the body uses plain-English labels and leaves the schedule trail in the methodology note.
As of April 23, 2026, the schedules, thresholds, and guidance referenced below are summarized from public FFIEC, FDIC, OCC, and Federal Reserve sources. Verify the latest filings and enforcement updates in the sources before citing this screen in a credit memo or investor document.
The snapshot date for the bank screen is December 31, 2025. It uses public Call Report data summarized in Banking Intelligence and should be validated against the FFIEC Central Data Repository public bulk data[1] and the FDIC December 2025 Call Report forms and instructions[2] before reuse. This screen is intentionally narrow: active U.S. banks with $1 billion to $50 billion in total assets, ROA above 1.00%, and NPL ratio above 2.50%.
The FDIC’s Fourth Quarter 2025 Quarterly Banking Profile[3] reported 1.24% ROA and $77.7 billion of aggregate quarterly net income for FDIC-insured institutions. That benchmark matters because this screen is not looking for weak earnings. It is looking for banks that still clear a profitability hurdle while also showing enough nonperforming-loan pressure to require a second look.
Key takeaways
- On the gap metric used here, MidFirst Bank is the clear outlier: a 14.85% NPL ratio, 1.19% ROA, and 12.5x NPL-to-ROA multiple.
- Barrington Bank & Trust Company, N.A. and Merrick Bank rank next by the same metric, but Merrick’s 11.40% charge-off rate makes it a different credit question than a high-NPL, low-charge-off bank.
- ROA above 1.00% and NPL ratio above 2.50% are internal screen cutoffs, not supervisory limits. They are useful for triage, but they break down when a bank has specialty consumer loans, unusual collateral, acquired portfolios, or a high-yield model.
- For fintech founders using this as a sponsor-bank check, pair the credit screen with the June 2023 Interagency Guidance on Third-Party Relationships[4] and the July 25, 2024 FIL-45-2024 bank-fintech statement[5].
- For bank boards and analysts, the next question is not whether high NPLs are automatically fatal. It is whether recurring pre-provision earnings, allowance coverage, collateral, and capital can absorb the credit risk visible in the public schedules.
The screen
Answer first: MidFirst Bank has the biggest gap in this screen by a wide margin, with an NPL-to-ROA multiple of 12.5x and almost no net charge-offs. Barrington Bank & Trust Company, N.A. and Merrick Bank are next by the same metric, but they require different reads because Merrick’s charge-off rate is already high. The caveat is important: the gap multiple is a triage ranking, not an expected-loss estimate, because ROA uses average assets while NPL ratio uses loans as the denominator.
Methodology: Gap multiple equals NPL ratio divided by ROA. ROA equals net income divided by average assets. NPL ratio equals loans 90 days or more past due and still accruing plus nonaccrual loans, divided by loans and leases. Charge-off rate is net charge-offs divided by loans and leases. The filters are active U.S. banks, $1 billion to $50 billion in assets, ROA above 1.00%, and NPL ratio above 2.50%. The ROA cutoff keeps the focus on banks that still clear a conventional profitability hurdle near the Q4 2025 industry ROA, while the NPL cutoff is an internal diligence threshold high enough to force a credit explanation. Both cutoffs need peer context by bank type.
| Bank | State | Assets | ROA | NPL ratio | Gap multiple | Charge-off rate |
|---|---|---|---|---|---|---|
| MidFirst Bank | OK | $41.4B | 1.19% | 14.85% | 12.5x | 0.01% |
| Barrington Bank & Trust Company, N.A. | IL | $4.8B | 1.04% | 5.39% | 5.2x | 0.05% |
| Merrick Bank | UT | $9.3B | 1.16% | 4.64% | 4.0x | 11.40% |
| Emigrant Bank | FL | $5.8B | 1.35% | 4.59% | 3.4x | 0.36% |
| International Bank of Commerce | TX | $4.4B | 2.18% | 6.28% | 2.9x | 0.12% |
| Comenity Capital Bank | UT | $13.6B | 1.60% | 4.19% | 2.6x | 8.18% |
| Comenity Bank | DE | $7.8B | 4.91% | 4.03% | 0.8x | 7.00% |
| Celtic Bank | UT | $4.8B | 4.12% | 2.81% | 0.7x | 0.76% |
The highest gap multiples in this screen belong to MidFirst Bank, Barrington Bank & Trust Company, N.A., Merrick Bank, Emigrant Bank, and International Bank of Commerce. That list is a triage queue, not a ranking of unsafe banks, because the table does not yet show collateral, allowance coverage, loan mix, borrower type, or whether the income is recurring.
Worked example: MidFirst Bank shows 1.19% ROA, 14.85% NPL ratio, and a 0.01% charge-off rate in the screen. Its 12.5x gap multiple does not mean NPLs are 12.5 times expected losses or capital need. It means visible problem-loan pressure is large relative to reported profitability, while realized charge-offs have barely appeared. The useful question is practical: why are nonperforming loans so high while charge-offs remain almost absent?
The next rows show why peer context matters. Barrington Bank & Trust Company, N.A. has a high gap multiple and low charge-off rate, while Merrick Bank has a high gap multiple and an 11.40% charge-off rate. Those are not the same credit story if loan composition, allowance coverage, and recurring revenue differ. Use the peer comparison view to compare banks with similar asset size and loan mix before assigning a risk narrative.
Why the gap happens
Profitability and asset quality diverge for specific, testable reasons. The point is to name the reason, tie it to a public filing line, and decide whether the explanation survives peer comparison.
- Specialty lending: A bank with a high-yield loan book can report strong ROA while also reporting high charge-offs. Comenity Bank’s 4.91% ROA and 7.00% charge-off rate should push the reader to interest income, net charge-offs and recoveries, and loan categories before comparing it with a plain commercial bank.
- Timing: Nonperforming loans can rise before losses are charged off. MidFirst Bank’s 14.85% NPL ratio and 0.01% charge-off rate make aging, charge-offs, and allowance movement the first three checks.
- Collateral and reserves: A high NPL ratio can be less severe if loans are well secured and the allowance is strong. The public-data test is allowance coverage: allowance for credit losses on loans and leases divided by the NPL numerator.
- Fee income and noninterest income: ROA can be supported by noninterest income even when loan performance weakens. Analysts should separate recurring fee revenue from one-time gains before treating ROA as durable loss-absorption capacity.
- Accounting and classification: The CECL model under ASU 2016-13, Topic 326[8] focuses on expected credit losses, while nonperforming status identifies past-due and nonaccrual loans. A nonperforming classification is not the same thing as a realized loss, and the allowance bridge matters.
The screen should therefore trigger diagnosis, not a one-word label. A high-profit, high-NPL bank can be earning through risk, reserving ahead of risk, or waiting for losses to emerge. The difference is visible only after ROA, NPLs, charge-offs, allowance coverage, capital, and loan mix are read together.
Follow-up checks
Use this workflow before calling a bank an outlier in a memo, board packet, diligence note, or article.
- Confirm the legal institution in FDIC BankFind[6], then pull the matching Call Report from the FFIEC CDR[7]. Do not rely on a bank name alone when there are affiliates, similarly named banks, or merger history.
- Recalculate ROA using net income and average assets. If ROA is above 1.00%, identify whether earnings are coming from net interest income, noninterest income, security gains, or other nonrecurring items.
- Recalculate the NPL ratio using loans 90 days or more past due and still accruing plus nonaccrual loans, divided by loans and leases. If the ratio is above 2.50%, move to loan mix before writing a conclusion.
- Compare NPL ratio with net charge-offs. A high NPL ratio with low charge-offs points to timing, collateral, modifications, or classification. A high NPL ratio with high charge-offs points to realized credit cost and capital absorption.
- Compute allowance coverage. Coverage below 100% is not automatically a problem, but it requires a clear explanation of collateral, guarantors, loan type, and expected loss assumptions.
- If commercial real estate is material, compare the bank with the December 2006 Interagency CRE Concentration Guidance[9]: construction, land development, and other land loans at 100% or more of total capital, or total CRE loans at 300% or more of total capital with 50% or more growth over the prior 36 months, are supervisory screening indicators, not hard legal limits.
- For sponsor-bank diligence, check public enforcement sources after the financial screen: the FDIC enforcement orders database[10], the OCC enforcement actions page[11], and the Federal Reserve enforcement actions page[12].
A simple decision table keeps the follow-up disciplined.
| Pattern | What it may mean | Next check |
|---|---|---|
| ROA > 1.00%, NPL > 2.50%, charge-offs < 0.25% | Problem loans are visible, but realized losses have not yet appeared in size. | Aging, allowance coverage, collateral, and loan type. |
| ROA > 1.00%, NPL > 2.50%, charge-offs > 5.00% | The bank may be earning through a high-yield, high-loss model. | Revenue durability, recoveries, capital ratios, and true peers. |
| NPL > 5.00% and ROA near 1.00% | Reported profitability may leave little margin if provision expense rises. | Quarter-over-quarter provision expense and allowance movement. |
| CRE concentrations near the 100% or 300% interagency screening indicators | Loan concentration can magnify the asset-quality signal. | Property type, geography, borrower concentration, stress testing, and board reporting. |
The underlying public-data context view is the right place to keep the schedule trail visible while you work through those checks. The goal is not to turn one ratio into a verdict. The goal is to make the next question specific enough that a banker, founder, journalist, or director can test it against the public record.
The practical takeaway
Use this decision rule tomorrow: keep a high-ROA, high-NPL bank on the candidate list only if three items are explainable in public data. Loan mix should explain why the NPL ratio is elevated, allowance coverage and collateral should explain loss absorption, and recurring pre-provision earnings should be strong enough to absorb charge-offs without pressuring capital.
If one of those items is missing, move the bank from “acceptable” to “diligence hold” until the next Call Report, management discussion, or peer comparison resolves it. For journalists, the rule is similar: do not write “strong bank” from ROA alone, and do not write “troubled bank” from NPL ratio alone. State the reporting date, the formulas, the denominator, and the missing follow-up item.
Sources
- FFIEC Central Data Repository public bulk data: https://cdr.ffiec.gov/public/PWS/DownloadBulkData.aspx
- FDIC December 2025 Call Report forms, instructions, and related materials: https://www.fdic.gov/bank-financial-reports/december-2025-call-report-forms-instructions-and-related-materials
- FDIC Fourth Quarter 2025 Quarterly Banking Profile: https://www.fdic.gov/quarterly-banking-profile/quarterly-banking-profile-q4-2025
- FDIC FIL-29-2023, Interagency Guidance on Third-Party Relationships: https://www.fdic.gov/news/financial-institution-letters/2023/fil23029.html
- FDIC FIL-45-2024, agencies issue statement on bank arrangements with third parties: https://www.fdic.gov/news/financial-institution-letters/2024/agencies-issue-statement-bank-arrangements-third-parties
- FDIC BankFind Suite for institution status and identifiers: https://banks.data.fdic.gov/bankfind-suite/
- FFIEC CDR facsimile view for Call Report filings: https://cdr.ffiec.gov/public/managefacsimiles.aspx
- Federal Reserve FAQ on ASU 2016-13 and current expected credit losses: https://www.federalreserve.gov/supervisionreg/topics/faq-new-accounting-standards-on-financial-instruments-credit-losses.htm
- Federal Reserve-hosted Interagency CRE Concentration Guidance: https://www.federalreserve.gov/frrs/guidance/interagency-guidance-on-concentrations-in-commercial-real-estate-lending-sound-risk-management-practices.htm
- FDIC enforcement orders database: https://orders.fdic.gov/s/
- OCC enforcement actions page: https://www.occ.gov/topics/laws-and-regulations/enforcement-actions/index-enforcement-actions.html
- Federal Reserve enforcement actions page: https://www.federalreserve.gov/supervisionreg/enforcementactions.htm