Success by JazE Edutech / Resource Kit
Resource Kit
An examiner's-eye kit pairing frequently flagged CRA data errors with prevention controls, a pre-submission audit checklist, and a correction-memo template.
The frequently flagged CRA data errors cluster in five families — geocoding mismatches, reportability misclassifications, assessment-area field gaps, loan-purpose coding errors, and revenue-size mistakes — and every one of them is preventable with a control that runs before submission, not after. This kit pairs each error family with its prevention control, then adds a pre-submission audit checklist and a correction-memo template so the file an examiner opens matches the lending the bank actually did.
Full written guide, sources, and FAQs
Run your CRA file through five inspection stations before an examiner does. This kit pairs each frequently flagged data error with a prevention control, a pre-submission audit checklist, and a correction-memo template.
This resource helps readers connect common CRA data errors to classroom practice, standards-aware implementation, and responsible next steps for schools and sponsors.
This kit is built for the people who own a bank's CRA data quality: CRA officers, compliance analysts, HMDA and CRA data stewards, loan-operations leads, and internal audit reviewers. If your signature is on a submission file — or on the certification behind one — this is your working tool. It is deliberately a quality-control layer, not another requirements explainer: it assumes the data-scoping work is done, and its job is to catch the errors that slip between capture and exam day.
It also pairs naturally with two companion resources. The Large-Bank CRA Data Readiness Kit scopes which records and fields the bank owes; the small-business reportability checker settles whether an individual loan belongs in the file. This kit takes over from there, testing, auditing, and correcting the assembled file until it can survive an examiner's eye. Banks of every charter type and size can walk the same five stations, scaled to the data fields their regulator expects.
Three working instruments, plus one shared frame. The frame is the Five-Station Inspection Line: a record-by-record mental model run like a production floor — every data point moves through capture, geocode, classify, complete, and certify, and the line returns one of three states — pass, flag, or fail. The instruments are an error taxonomy with matching prevention controls, a pre-submission audit checklist, and a correction-memo template for anything found late.
Take one real record — say a $180,000 small-business term loan booked in March — and walk it down the line in station order. At Station 1, Capture, the record must exist in the reporting extract with the field set the bank owes. At Station 2, Geocode, the address must resolve to a census tract you can defend. Station 3, Classify, tests reportability. Station 4, Complete, screens every required field for blanks, out-of-range values, and codes that contradict the loan file. Station 5, Certify, is the signed result of the full-file audit.
Each station returns a visible state. Pass moves the record forward untouched. A flag marks a judgment call — a remodeled property, a split-purpose loan — that a second reviewer must resolve. A fail stops the line: the record drops down the fail-out chute, enters the correction-memo process, and does not return until the fix is documented. Because the states are visible, a file review becomes a diagnostic — count the flags at each station and you know exactly which control needs reinforcement before examiners count the same errors from the other side of the table.
Examiners do not grade data quality in the abstract — they test whether the file tells the truth about where and to whom the bank lent. That is why the recurring errors cluster into five families, and why each family has a natural control point. None of these controls is exotic; they are timing and ownership decisions made early enough to be cheap.
Data integrity is part of the exam itself: the CRA regulations at 12 CFR Part 345 set out the data and recordkeeping duties a bank owes its primary federal regulator, and the agencies' CRA examination procedures direct examiners to review the accuracy and completeness of reported loan data — so a file with systematic errors invites reviewers to dig further into records that were otherwise fine. The purpose of this taxonomy is a boring file — one where the story the data tells and the story the loan files tell are finally the same story.
Run this as a two-person pass: a preparer works the list, a reviewer signs it, and neither may be the only set of eyes the file gets. The checklist is deliberately short enough to finish — a checklist that takes a week gets skipped, and a skipped checklist is how systematic errors survive to exam day. Log every exception with an owner and a deadline, even the ones you fix on the spot.
A correction without a memo is a correction an examiner has to rediscover. The template is one page: the record and field that was wrong; when and how the error entered the data; the correction applied and its effective date; the control added so the same error cannot recur; and the name of the person who approved the fix. Memos turn late-found errors from liabilities into evidence that the quality-control system works.
Position matters: this kit is the quality-control layer between data scoping and exam day. Scope requirements with the Large-Bank CRA Data Readiness Kit, settle individual loan classifications with the small-business reportability checker, and when the process outgrows spreadsheets, use the CRA platform selection scorecard to choose tooling that automates these stations. Together the resources walk the full lifecycle: know what you owe, classify each loan, quality-check the file, then pick the platform.
This kit is an educational planning resource, not legal, compliance, supervisory, or regulatory advice, and it does not create any expectation about examination outcomes. Data requirements differ by bank category, charter, and primary federal regulator, and they change: use your regulator's current rules, examination procedures, and reporting instructions as the controlling documents, and involve your compliance counsel before changing any submission practice.
How any specific data correction, filing, or activity is treated depends on program facts and examiner review, and no checklist can substitute for that review. Nothing here promises a CRA rating outcome, regulatory acceptance, or supervisory credit. When this kit and a regulator's current guidance disagree, follow the regulator.
The recurring errors cluster in five families: geocoding mismatches, reportability misclassifications, assessment-area field gaps, loan-purpose coding errors, and revenue-size mistakes. Agency examination procedures direct examiners to review the accuracy and completeness of reported loan data, so systematic errors in any of these families invite a closer look at the records behind them.
Most geocoding errors start upstream of the data team: post-office boxes and rural-route addresses that never resolve cleanly, properties on annexed or renumbered streets, stale address standards in origination systems, and census tract boundary changes after each decennial census. The control is timing — geocode at booking through the FFIEC Geocoding System, then re-verify the highest-risk slices every cycle.
Errors are found and corrected all the time; what examiners weigh is pattern and response. Isolated mistakes fixed through a documented process read very differently from systematic errors discovered without a quality-control trail. Banks that find their own errors, correct them, and can show the control that prevents recurrence generally walk into exams with a stronger position — though outcomes always depend on the facts and the examiner's review.
A workable rhythm is continuous controls at the point of capture, a sampling pass each reporting cycle, and the full two-person audit checklist before every submission. The right frequency scales with volume and change: a bank mid-conversion, mid-acquisition, or working through a census boundary update should sample more, not less, until the file stabilizes.
Six things: the record and field that was wrong; when the error entered the data; the root cause; the correction applied and its effective date; the control added to prevent recurrence; and the approver's name. One page is enough — the memo exists so the fix can be verified without a meeting.
The controls scale, but the discipline does not. Data requirements differ by bank size and category under each regulator's CRA rules, so a smaller bank may owe fewer fields — yet every field it does owe faces the same question an examiner asks of a large bank: does the file tell the truth? Right-size the sample, not the standard.
The 2023 CRA final rule widened the set of data banks report, which widens the surface these same errors can appear on. The Five-Station Inspection Line absorbs new fields without changing shape: each new field still needs a defined source at capture, a completeness screen, and a place in the sampling plan. Compliance dates have shifted since the rule was issued, so check your regulator's current timeline for what applies to your bank and when.
Federal Financial Institutions Examination Council
Board of Governors of the Federal Reserve System
Federal Deposit Insurance Corporation
Office of the Comptroller of the Currency
Office of the Federal Register and GPO
Federal Financial Institutions Examination Council
Federal Financial Institutions Examination Council
Walk one small business loan through the reportability gates: product, size, and borrower revenue decide whether it must be reported for CRA. See which thresholds apply and which data points the file should capture.
A control-tower-style readiness kit that turns the large-bank CRA data collection rules into field inventories, a two-date compliance runway, and a pre-submission audit your team can actually run.
A decision worksheet, exam-file checklist, and note templates that walk a single PPP loan file to one of two CRA outcomes — reported small business lending or documented community development financing.