
Below the Line Testing in BSA/AML
The transaction-level half of independent testing: what a tester samples, what the file has to prove, and how to be ready before they ask.
The short answer
Below the line testing is the transaction-level part of an independent BSA/AML review. Instead of asking whether your program exists on paper, the tester pulls actual customer records and transactions and checks whether your controls worked in practice: whether currency transaction reports were filed correctly and on time, whether suspicious activity was escalated and reported when it should have been, and whether customer due diligence files contain what your policy says they contain. It is the counterpart to above-the-line testing, which reviews the program's design, policies and governance.
Above the line vs. below the line
The split is between asking whether the program is designed correctly and asking whether it actually ran. Most review scopes contain some of each, and a report that only does the first half leaves the harder question unanswered.
Above the line
Design and documentation
- The written AML program and its procedures
- The BSA officer's role, authority and reporting line
- The documented risk assessment
- The training program and its completion records
- Management oversight and escalation paths
Below the line
Operating effectiveness, using real data
- Transaction testing against source records
- Filing accuracy and timeliness for CTRs and SARs
- Alert disposition and case file quality
- Customer due diligence and ownership files
- Whether exceptions were escalated and resolved
Both are needed because they fail differently. A policy can look complete while the controls underneath it quietly stop working: a report gets filed late, an alert gets closed with a one-line note, a certification goes stale. Only transaction-level testing shows which situation you are actually in. Our guide to independent AML testing covers the review as a whole, including how the two halves fit into one scope.
What below the line testing typically covers
Scope varies by business type, but the same six areas come up across money services businesses, non-bank lenders, fintechs, title and escrow companies and real estate businesses.
Currency transaction reporting
Whether reportable transactions were identified, aggregated correctly across the business day, and filed within the required window rather than caught later and filed late.
Suspicious activity reporting
Whether alerts and internal referrals led to a documented decision, and whether a filing was actually made when the facts in the file supported one.
Customer due diligence and beneficial ownership
Whether required identifying information and ownership certifications are present, current and verified, not merely collected once at onboarding and left to age.
Transaction monitoring
Whether alerts were reviewed, escalated or closed with a reasoned narrative, and whether the thresholds still fit the activity the business actually generates.
OFAC and sanctions screening
Whether customers and transactions were screened at the right moments and whether potential matches were resolved with a documented conclusion.
Recordkeeping
Whether the records your program promises to keep are genuinely retained, retrievable and complete when someone asks for them.
The terminology shifts from firm to firm. If a term in a scope document is unclear, the AML/BSA glossary covers the common ones in plain language.
How a tester builds the sample
There is no single required method, but the process follows the same arc in almost every engagement. Sample size and selection are judgment calls, and the tester should be able to explain both.
Define the population
The tester starts with the full set of transactions or customers that fall inside the test period and inside the scope, for example every cash transaction above the reporting threshold, or every alert raised by monitoring. If the population is incomplete, everything after it is unreliable.
Select the sample
Items are then chosen from that population. Selection can be risk-based and judgmental (the largest transactions, the highest-risk customers, the alerts closed fastest) or statistical. Both are legitimate; what matters is that the tester can explain why that selection answers the question being tested.
Test each item against your policy and the rule
Each sampled item is checked twice: did the business do what its own written procedure says, and does that procedure meet the applicable requirement. The gap between those two answers is where most useful findings come from.
Document exceptions
Every item that fails is written up with enough detail that a reader who never saw the transaction can follow what happened, what should have happened, and what evidence supports the conclusion.
Report by severity
Findings are rated rather than listed flat, because a late filing on a single small transaction and a pattern of missed filings are not the same problem. Ratings tell the BSA officer where remediation effort belongs first.
Where it fits in the five BSA/AML pillars
Independent testing is one of the five pillars of a BSA/AML program. Below-the-line work is the part of that pillar which verifies the other four actually operate in daily practice.
Written policies and procedures
A designated BSA compliance officer
Ongoing employee training
Independent testing
Customer due diligence
What to have ready
A tester will ask for evidence, not assurances. Having these assembled and organized before the request arrives shortens the engagement and removes the impression that something is being hidden.
Transaction logs and CTR and SAR filing records for the test period, with filing dates visible
Alert and case files with dispositions, escalation notes and closure rationale
Customer due diligence and beneficial ownership files for the customers in scope
OFAC and sanctions screening logs with match resolutions
The written program, the current risk assessment and training completion records
Prior findings and the evidence that each one was remediated
Be ready to explain exceptions, not just hand over documents. A tester who hears "that one was filed late because our alert queue backed up in March and we caught it in April" is looking at a business that understands its own control failure. A tester who receives a stack of PDFs and no explanation learns less, and writes the finding the documents suggest.
Two practical starting points before an engagement: run through the AML compliance checklist to confirm the documentation exists, and use the free BSA/AML audit simulator to see how your own answers read when someone else scores them.
Common findings
These come up repeatedly, and most of them are process problems rather than bad faith. That is exactly why they are worth looking for before someone else does.
CTRs filed late or missed for aggregated activity
Multiple cash transactions that together crossed the threshold were never combined, so no report was filed, or the report went out well after the filing window closed.
SAR decisions with no documented rationale
A case was closed as "no SAR required" with nothing in the file explaining what facts were considered or why they did not rise to reporting.
Beneficial ownership certifications missing or stale
Ownership certifications were collected at onboarding and never refreshed, or a legal entity customer has no certification in the file at all.
Alerts closed with one-line notes
"Reviewed, no concern" tells a reviewer nothing. It cannot be tested, defended or relied on later.
A review performed by someone who is not independent
When the BSA officer, or an employee who reports to the BSA officer, performs the review, the work may be sound and still fail the independence expectation.
Questions to ask a prospective tester
Two proposals can both say "independent testing" and describe very different work. Ask these six questions and compare the answers, not the cover page.
Does your scope include transaction testing, or only a policy review?
A document review and a transaction test produce very different reports. Ask which one you are buying, in writing, before the engagement starts.
Which populations will you test, and how will you select the sample?
The answer should name the specific populations (CTRs, alerts, CDD files, screening hits) and the selection logic, not just "a sample of transactions."
Will you test CTR and SAR filings against the underlying source records?
Re-reading a filing proves the form was submitted. Rebuilding it from the transaction proves the filing was right.
How do you document and rate exceptions?
You want to see a sample finding write-up. If the report only counts exceptions, you cannot prioritize remediation.
Will the report map findings to the five pillars?
A pillar mapping is what lets you hand the report to a board, a banking partner or an examiner without rewriting it.
Who performs the work, and is that person independent of the program?
Independence is the point of the exercise. Confirm who does the testing and who they report to inside their own firm.
How Soflo Consulting helps
Soflo Consulting is an online AML/BSA compliance platform for regulated businesses across all 50 states. Training plans include annual awareness training with certificates, review of an existing AML/BSA program, creation of an AML/BSA program, customized training modules, and banking-partner readiness support.
Soflo does not perform independent testing. Independent testing has to be performed by a party independent of the program. That does not require an outside firm: a larger institution can use internal audit, and a smaller business can use an employee who is neither the BSA officer nor someone reporting to that officer.
Soflo's role is on the other side of that line. The work we do leaves clients with the written program, the risk assessment, the training records and the supporting documentation an independent tester will ask for, so the testing engagement is about whether the controls ran, not about whether the paperwork exists. If the risk assessment itself is the weak point, that is what our BSA risk assessment support is for.
Get the documentation in place before the tester asks
Written program, current risk assessment, training records with certificates, and supporting evidence, organized so an independent review can move straight to transaction testing.