Research · Published:
Who owns an unresolved vendor credit balance? An evidence review
A bounded study of credit origins, applications, refunds, entity boundaries, and handoff ownership in outsourced AP reconciliation.
Research question and method
Who should own an unresolved vendor credit balance at each stage, and what evidence lets accounts payable distinguish a usable credit from an unexplained negative amount? This research follows credits from source document or payment event through ledger entry, supplier statement, proposed application, refund, and closure. GAO internal-control guidance informs documentation and responsibility. NIST guidance informs access to supplier, payment, and accounting records. IRS Form W-9 material is included only as context for controlled supplier identity fields, not as authority over credit treatment. The sample contains a credit memo, overpayment, returned goods, duplicate-payment recovery, cross-entity balance, disputed deduction, promised refund, and stale open credit. The aim is to test evidence and ownership, not to prescribe accounting or collection policy.
Evidence population
Each sampled balance needs a declared entity, supplier identifier, currency, source type, amount, original invoice or payment reference, source date, posting record, statement line, communications, proposed disposition, current owner, and cutoff. A ledger balance is a system fact at a stated extraction time. The statement may show the supplier's position. Calling the two amounts the same credit is analysis until identifiers and history support the relationship. Applying the balance to an invoice, requesting a refund, writing it off, or transferring it between entities is a decision. Outsourced AP support can place the sources side by side and describe the relationship without making that decision.
Case analysis
The cases should resist easy netting. A supplier statement may combine a credit with invoices for another legal entity. A cash refund may arrive without a reference and remain unapplied. A replacement invoice may reduce the original charge while a separate credit memo also appears. The study retains gross events and links them; it does not hide uncertainty by forcing a zero balance. For every proposed application, show who proposed it, the source used, which invoice would change, and which employee accepted or rejected the proposal. If the supplier confirms a balance by email, preserve the message and the controlled supplier identity used to evaluate it. Confirmation is evidence of the supplier's statement, not automatic authority to post.
Reproduction test
Reproduction starts from the open-credit report at a fixed cutoff. A reviewer traces the balance back to its originating record and forward to every application, reversal, refund, or dispute event. A second reviewer repeats the trace without an oral briefing. Include cases where references are incomplete, currencies differ, records were corrected, and a supplier relationship spans entities. Record excluded systems and inaccessible documents. When reviewers produce different remaining balances, keep both bridges and locate the disputed event or rule. A later posting must be appended to the cutoff view rather than changing what the first review saw. This preserves the difference between an aging snapshot and the balance eventually resolved.
Operating boundary
Ownership changes with the question. AP support may collect credit documents, reconcile statements, maintain an open-item register, send approved factual follow-ups, and prepare an application proposal. Procurement may own a return dispute. Treasury may own confirmation of cash received. Accounting decides posting, write-off, reclassification, and period treatment. An authorized finance owner approves application and closure under company policy. The queue should therefore identify both the evidence task and the decision owner. Assigning the whole balance to "AP" conceals which action is actually pending and can leave outsourced staff appearing responsible for decisions they cannot make.
Interpretation
The analysis uses state transitions rather than a single resolved flag. Suggested descriptive states are source missing, identity or entity conflict, supplier confirmation pending, application proposed, employee decision pending, refund promised, cash receipt unlinked, accounting entry pending, and closed with retained disposition. These are research categories, not universal workflow requirements. Aging should begin from a declared event for each state. A promised refund is not recovered cash. A posted application is not proof that the supplier agrees. A zero ledger balance is not proof that the underlying dispute is closed. Preserving those distinctions prevents a neat report from overstating evidence.
Limitations
Access controls matter because credit work touches invoices, bank events, supplier tax records, and ledger entries. Support should receive only the views and preparation functions needed for the assigned population. It should not edit vendor-master identity, create refunds, post journals, approve write-offs, or release payments. A system role that combines reconciliation and posting should be treated as broader access even if staff are instructed not to use it. The study records the technical capability, written role, and review control separately. Where a source cannot be shown because of permission limits, the packet names an authorized reviewer rather than copying sensitive data into an uncontrolled file.
Evidence-led conclusion
Limitations: public control sources do not determine whether a credit is legally enforceable, how it should be recognized, when it becomes stale, or whether it can cross entities. Supplier statements may omit disputes or have different cutoffs. Ledger extracts may exclude unposted items. A bounded sample cannot estimate recoverable value or staff productivity, and selected aging items may overrepresent difficult cases. Currency conversion, tax consequences, escheatment, and contractual rights fall outside this research. The company should have accountable owners decide those matters. The evidence test only shows whether the sampled balance and its next decision can be reconstructed.
Findings
Evidence-led conclusion: an unresolved vendor credit becomes governable when its origin, entity, currency, applications, reversals, communications, and present decision owner remain linked at a stated cutoff. Outsourced AP support can maintain this chain and surface the exact missing action. It should not turn a proposed application, supplier promise, or unexplained negative balance into a closed result. The best handoff states what the sources prove, what the comparison suggests, what remains unknown, and which authorized employee can decide it. That structure gives finance owners a useful queue while keeping accounting and cash decisions inside the company.
Sources
These primary sources support the control principles and evidence boundaries in this report.
FAQs
Are the planning numbers benchmarks?
No. They describe a testable workflow shape and are not promises, market averages, or production targets.
What should an outsourced AP assistant own?
Repeatable preparation, documentation, status tracking, and follow-up within least-privilege access. Named finance owners retain approval and payment decisions.
When should an item be escalated?
When evidence is missing, a request changes payment details, a duplicate or fraud signal appears, or the item falls outside the written rule.