Accounts Payable Reconciliation: Tying the AP Subledger to the GL Control Account

The Control That Proves the Number Is Real

The accounts payable figure on the balance sheet is a single line. Behind it sits the AP subledger: thousands of individual vendor balances, each supported by invoices, receipts and payment records. The reconciliation between the two is what proves the reported number means something.

When the subledger total equals the GL control account, every dollar of reported liability traces to a specific vendor and a specific invoice. When it does not, the difference is unsupported. Something is in the reported liability that no vendor is owed, or something is owed that the reported liability does not include. Both misstate the balance sheet, and completeness of liabilities is consistently one of the assertions auditors weight most heavily, because understating payables flows straight through to overstated profit.

This is distinct from the other reconciliations in a typical close. Bank statement reconciliation compares your cash record to the bank’s. Vendor statement reconciliation compares your vendor balance to the supplier’s own record. Both are external controls testing agreement with a third party. Subledger-to-GL reconciliation is internal: it compares two of your own records and tests whether your own systems are internally consistent. A team can pass vendor statement reconciliation for every supplier and still have a broken control account, because the two tests look at entirely different things.

The Six Causes of Drift

When the two balances disagree, the cause is nearly always one of six things. Working through them in order of likelihood turns a day-long hunt into a structured hour.

Manual journals posted directly to the control account. This is the dominant cause and the most preventable. Someone posts an adjusting entry straight to the AP control account, often to correct an expense classification or to book an accrual. The entry hits the ledger but has no vendor-level counterpart, so the subledger cannot move with it. The difference is permanent until reversed. Most ERPs can lock the control account to subledger postings only, and enabling that single setting removes more reconciliation pain than any other intervention.

Period and posting-date differences. An invoice is posted to the subledger in the new period while the GL entry lands in the old one, or vice versa. These are legitimate reconciling items, but they must be listed and cleared, not absorbed.

Foreign currency revaluation mismatch. In multi-currency environments, the subledger and the GL must be revalued at the same rate on the same date. When revaluation logic diverges, the difference recurs every month and grows as rates move. Teams often misread this as a systemic error when it is a configuration issue, a pattern we see frequently in multi-currency AP reconciliation.

Multiple control accounts and misposting. Organisations with separate control accounts for trade payables, intercompany payables and accrued liabilities frequently have invoices posted against the wrong one. The consolidated total may look correct while both individual accounts are wrong, which is why reconciliation must be performed per control account rather than in aggregate.

Interface and integration failures. Where AP runs in one system and the GL in another, a partially failed batch posts some transactions and not others. This is the most dangerous cause because it is silent; nothing errors, the balances simply diverge. Organisations on file-based rather than API integration are especially exposed, as covered in our comparison of file-based versus API ERP integration.

Accruals and GRNI hitting the AP control account. If unbilled receipts post to the AP control account rather than a separate accrual account, the ledger includes liabilities that have no vendor-level subledger entry, because no invoice exists yet. Keeping GRNI in its own clearing account is the correct structural answer.

Isolating a Difference in Under an Hour

When a variance appears, the instinct is to start scanning transactions. That is the slowest possible approach. A structured narrowing sequence is far faster.

Begin with the size and shape of the difference, because it is diagnostic on its own. A difference that is a round number is almost always a manual journal. A difference that changes every month in proportion to FX rates is a revaluation issue. A difference that equals the exact value of one invoice is a misposting. A difference that has been constant for several periods is an old unreversed entry that nobody has cleared.

Then compare transaction counts, not just balances. Two systems can show the same total while holding different transaction populations, particularly after a partially failed interface batch. Count mismatch with matching totals is a strong signal of an integration problem.

Next, filter the GL control account for entries with no subledger reference. In most ERPs this is a single query and it resolves the majority of differences immediately.

Only then move to vendor-level comparison, and do it by exception rather than exhaustively. Compare subledger vendor balances to the sum of GL entries by vendor where the source system supports it, and investigate only the vendors that differ.

Unrecorded Liabilities: The Reconciliation That Balances but Is Still Wrong

A subtlety worth stating plainly: a perfectly tied reconciliation does not mean liabilities are complete. The subledger can agree exactly with the GL while both omit obligations the business genuinely incurred before period end.

This is the completeness problem, and it is why auditors perform a search for unrecorded liabilities separately from the tie-out. The standard technique reviews invoices and payments processed in the weeks after period end, identifying any that relate to goods or services received before the cut-off. Those should have been accrued and were not.

The three populations worth reviewing every period are post-close invoices relating to pre-close delivery, aged GRNI where receipts exist without invoices, and open purchase orders with confirmed receipt and no invoice. A team with disciplined GRNI ageing has already covered most of this exposure, which is one reason the two controls are best operated together rather than as separate close tasks.

Guidance on liability recognition and completeness is set out by the IFRS Foundation, with practical interpretation available from IAS Plus and the IFAC. Reconciliation control expectations are covered well by Corporate Finance Institute and in close-process research from Deloitte.

How Peakflo Helps

Peakflo automates the mechanical portion of AP reconciliation, which is nearly all of it. Balances are extracted from the subledger and the GL control account at a synchronised cut-off, the variance is calculated automatically, and entries posted to the control account without a subledger source are flagged immediately rather than discovered through manual scanning. Transaction counts are compared alongside totals, so a partially failed interface batch surfaces as an exception instead of drifting silently into the next period.

Reconciling items are documented with supporting evidence and expected clearing dates, producing an audit-ready working paper without manual assembly, and the same engine ages GRNI so completeness exposure is visible in the same view as the tie-out. With native connectors to systems including NetSuite and Microsoft Dynamics 365 Business Central, and with payments reconciliation running on the same platform as invoice processing, reconciliation becomes a monitored control rather than a close-week task. To see it applied to your own ledgers, request a demo.

Our Verdict: Fix the Cause, Not the Month

After analysing why AP reconciliations consume close capacity, here is our recommendation.

Treat reconciliation as a priority problem if

• The same difference recurs every month and is re-investigated each time
• Any unexplained residual is carried forward rather than cleared
• Reconciliation takes more than half a day per entity
• Auditors have raised control account or liability completeness findings
• Manual journals to the AP control account are permitted by configuration

Automate the tie-out when

• You reconcile more than three entities, where effort scales linearly
• Multi-currency revaluation is in scope
• AP and the GL sit in different systems connected by an interface
• Close timelines are compressed and reconciliation sits on the critical path

Do not rely on the tie-out alone if

• GRNI ageing is not separately controlled, since a balanced reconciliation can still omit real liabilities
• Vendor statements are not reconciled, since internal consistency does not prove external agreement

Our Recommendation: Before optimising the reconciliation process, close the two configuration gaps that cause most of the work. Lock the AP control account so it accepts subledger postings only, and move accruals and GRNI to separate accounts. Those two changes eliminate the majority of recurring differences outright, which is a better outcome than automating the investigation of differences that should never occur.

Conclusion

AP subledger-to-GL reconciliation has an unusual property among close tasks: when the underlying controls are right, it is almost trivial, and when they are wrong, it is close to impossible. There is very little middle ground. Teams spending three days per entity are not doing a harder version of the same task as teams spending twenty minutes; they are investigating structural drift that should have been prevented at configuration level.

That is the shift worth making. Rather than resourcing the monthly investigation better, ask which of the six causes is generating the difference and remove it. A locked control account, separated accrual accounts, aligned FX revaluation and a monitored interface between systems together eliminate most recurring variances permanently.

And remember that a tied reconciliation is necessary but not sufficient. The subledger and the ledger can agree perfectly while both understate what the business actually owes. Pair the tie-out with disciplined GRNI ageing and a post-close invoice review, and the payables figure becomes genuinely defensible rather than merely internally consistent.

Frequently Asked Questions

What is accounts payable reconciliation?

Accounts payable reconciliation is the control that confirms the total of the AP subledger, meaning the sum of all open vendor balances, equals the accounts payable control account in the general ledger. It proves that reported liabilities are supported by actual vendor-level detail and that no unsupported entries have been posted to the control account.

How do you do an accounts payable reconciliation?

Extract the AP ageing summary and the GL control account balance at the same cut-off, compare the totals, then isolate the difference by checking for manual journals posted directly to the control account, timing and posting-period differences, foreign currency revaluation mismatches, and misposted vendor transactions. Document every reconciling item with evidence and correct the unexplained residual.

What is the difference between AP reconciliation and vendor statement reconciliation?

AP subledger-to-GL reconciliation is an internal control comparing two of your own records to prove internal consistency. Vendor statement reconciliation is an external control comparing your vendor balance against the supplier’s own record of what you owe. They test different things, and passing one says nothing about the other.

Why does the AP subledger not match the general ledger?

The most common causes are manual journal entries posted directly to the AP control account, transactions posted to a different period in one system than the other, foreign currency revaluation applied inconsistently, invoices posted to the wrong control account when multiple AP accounts exist, incomplete or failed interfaces between systems, and GRNI or accrual entries hitting the AP control account rather than a separate accrual account.

Should the AP subledger always equal the GL exactly?

Yes, after legitimate reconciling items are listed. A properly performed reconciliation may show timing differences such as an in-transit payment, but each must be explained and supported. An unexplained residual, however small, indicates a control weakness because the same mechanism producing a small difference can produce a large one.

Can you post journal entries directly to the AP control account?

You generally should not. Direct journals to the AP control account break the link between the subledger and the ledger by definition, because the entry has no vendor-level counterpart. Most ERPs allow the control account to be locked to subledger postings only, and enabling that setting eliminates the largest single source of reconciliation drift.

How long should an AP reconciliation take?

With clean controls and automated extraction, a single-entity reconciliation should take under 30 minutes and is often fully automated. Teams routinely spending one to three days per entity are usually investigating recurring structural drift rather than performing a reconciliation, which means the underlying cause has never been fixed.

What are unrecorded liabilities and how do you find them?

Unrecorded liabilities are obligations incurred before period end that were not posted, so the balance sheet understates what is owed. The standard search technique reviews invoices and payments processed shortly after period end to identify any relating to goods or services received before the cut-off, alongside the aged GRNI balance and open purchase orders with receipts but no invoice.

Why do auditors test AP reconciliation?

Auditors test it because completeness of liabilities is one of the highest-risk financial statement assertions, and understating payables directly overstates profit and net assets. The subledger-to-GL tie-out is the primary evidence that the reported payables figure is supported by real vendor-level obligations.

How does multi-entity structure complicate AP reconciliation?

Each entity requires its own tie-out, and intercompany payables must eliminate cleanly on consolidation. Difficulty multiplies when entities run different ERPs or charts of accounts, because a difference may sit in the interface between systems rather than in either ledger, and because intercompany balances often disagree between the two sides.

Can AP subledger to GL reconciliation be automated?

Yes. Balance extraction, variance calculation, identification of direct journals to the control account, timing-difference detection and reconciliation documentation can all be automated, leaving humans to investigate only genuine unexplained residuals. This is one of the highest-value close automations because the work is rule-based and repeats identically every period.

What should an AP reconciliation working paper contain?

It should show the GL control account balance, the AP subledger total, the variance, an itemised list of reconciling items with supporting evidence and expected clearing date, the unexplained residual, and dated preparer and reviewer sign-off. Auditors assess the quality of the reconciling item explanations as closely as the arithmetic.