How Should a Finance Team Test Month-End Close Controls and Find Financial Discrepancies?

What Month-End Close Testing Actually Means

Month-end close testing is the repeatable process of examining whether a company’s closing procedures produce complete, accurate, timely, and independently verifiable financial information. It is not a single software feature, a one-month forensic audit, or merely balancing the general ledger. Testing may cover journal entries, bank reconciliations, account reconciliations, intercompany transactions, accruals, estimates, consolidation, financial statements, and close-related system permissions. The objective is to identify discrepancies before they become recurring errors, misleading reports, or late regulatory filings. Evidence should be re-created rather than accepted merely because the preparer says a balance was checked. For context, one widely cited 2026 accounting-AI discussion reported that even the best tested model still failed one in five accounting tasks, so automated assistance should augment—not replace—review by qualified finance staff. A sound program turns a successful close into evidence that the process itself operated consistently.

Also worth reading: How Should AI Audit Evidence Controls Strengthen Financial Audits in 2026? · What Are the Best AI Model Risk Controls for Financial Services in 2026? · How Do Companies Optimize Internal Financial Controls Without Slowing Down the Business?

Organizations should distinguish three different forms of review. A control test asks whether a defined procedure was performed by the right person at the stated frequency and with the required evidence. A substantive review asks whether account balances and transactions are materially correct, including unusual items and estimates. A financial statement audit examines whether the statements comply with the applicable framework and whether material misstatements exist. A team can test controls without being engaged to audit financial statements, and a clean control sample does not guarantee that every balance is correct. Month-end close testing is most effective when both procedure design and the resulting numbers are evaluated. The reporting period, material accounts, risk level, and applicable reporting rules determine its depth.

Why Month-End Testing Finds Discrepancies Others Miss

Most financial discrepancies arise not from random arithmetic failure but from weak handoffs, duplicated entries, omitted accruals, incorrect cutoffs, stale master data, or unsupported judgment. A payment recorded one day late can distort cash, liabilities, income, and several ratios at the same time. An accrual reversed for the wrong reason can be re-expensed in the next period, creating a recurring expense pattern rather than a visible one-time error. Bank reconciling items left unresolved for 30, 60, or 90 days are particularly useful warning signs because they indicate a control failure even when the net cash balance is correct. Testing from transaction to ledger, and from ledger back to source evidence, catches errors that a one-directional review misses.

Control design is therefore more important than the sophistication of the close calendar. A close completed in five days can still be unreliable if management overrides approvals, spreadsheets lack change control, or consolidation extracts omit a subsidiary. Conversely, a complex review may be justified for a public company with multiple reporting deadlines and consolidation obligations, while a small cash-basis business may need a much lighter approach. Public-company accelerated filers commonly have tighter SEC deadlines—60 or 75 days depending on filer status—than non-accelerated filers, who generally have 90 days for annual Form 10-K reports and 45 days for quarterly Form 10-Q reports. The significance of a delay depends on the filing obligation, not only the number of days management selected internally.

Testing also examines whether a discrepancy is isolated or systemic. One unsupported manual journal may indicate a staff training problem, while a recurring year-end adjustment of similar size may point to faulty accrual methodology. Error rates should be tracked by control, process owner, source system, and root cause. A finance leader should not claim that close quality improved merely because fewer closing tasks were marked complete; the relevant measures are supported items, timely reconciliations, correction time, repeat exceptions, and material adjustments. This changes the activity from ceremonial sign-off into measurable process control.

A Practical Testing Method for Each Month-End Close

Begin by documenting the close calendar, including hard reporting dates, feed dates, review deadlines, and dependencies. Identify the population of journal entries, manual adjustments, bank reconciliations, account reconciliations, consolidation packages, and system-access changes. Testing can cover all material accounts on a judgmental basis, a risk-based sample, or the entire population when software permits inexpensive automated matching. Each test should state its objective, population, frequency, evidence standard, exception threshold, reviewer, and remediation owner. For example, testing might require a signed reconciliation, a statement dated no earlier than the bank’s reporting date, and documented support for items outstanding more than 10 business days.

Re-performance is the strongest way to validate a recurring calculation. Start with the source report, reperform additions and cutoffs, trace selected transactions to the general ledger, and agree subtotals to the consolidation system. Journal-entry testing should examine manual and top-side entries rather than selecting randomly from thousands of system-generated transactions. A useful filter includes users with posting privileges, unusual amounts, entries posted near period end, entries with round-dollar amounts, entries that increase revenue or reduce expense, and entries posted outside normal business hours. Companies frequently apply a materiality screen, but analysts should also investigate patterns below that threshold because many small exceptions can create a material aggregate effect.

Evidence should be retained in a controlled repository with the period, preparer, reviewer, completion timestamp, version, and exception disposition. Emails saying “reviewed” are weaker than a dated reconciliation showing the bank statement, book balance, identified items, adjusted balance, and reviewer approval. Exceptions should be assigned corrective actions and retested, not simply cleared by changing the status field. A common cadence is interim testing during the first close, expanded testing during the second close, and a limited year-end validation of mature controls. The exact number of test cycles depends on organizational change, turnover, acquisitions, new systems, and the reliability of earlier evidence.

Comparing the Main Month-End Testing Approaches

No single alternative suits every finance organization. Manual testing offers flexibility but consumes reviewer time and creates inconsistent documentation. Rule-based automation provides repeatable population matching, yet it can produce false alerts when data quality is weak. Continuous controls monitoring detects changes throughout the month, while a traditional month-end review is easier to organize but may discover issues after transactions have affected several periods. External audit is not a substitute for management’s close controls because management remains responsible for the financial statements and internal control process.

Manual testing is appropriate when transaction volumes are modest, the process depends on professional judgment, or automation tools have not been validated. Automated monitoring is usually more practical where one feed feeds several ledgers, account populations are large, or a failed control must be detected daily. Some organizations use continuous monitoring during the month and retain a smaller end-of-month testing set, which reduces the effort required after the ledger is already locked. External audit should test relevant controls when necessary, but it should not be counted as management operating an effective close control every month. Selection should follow risk rather than product popularity.

Journal Entries, Reconciliations, and Financial Statement Tests

Journal-entry testing begins with an independently generated population, not a report filtered by the person being tested. Reviewers should determine whether the entry has business purpose, authorization, support, correct accounting period, appropriate tax treatment, and correct entity or account coding. A technically valid debit and credit can still be financially wrong if the expense belongs to the prior period or was already recorded. Recurring entries should be compared with the original accrual and with the next-period reversal. Entries prepared by senior management warrant particular attention because authority does not eliminate the need for evidence, and management override is a recognized audit risk.

Bank reconciliation testing should verify the statement version, book balance, reconciling items, arithmetic, and disposition of stale items. Cash accounts with unusual negative balances, large transfers between entities, or a high volume of round-dollar journal entries merit expanded review. Accounts receivable and inventory require more than agreement of the trial balance. The age distribution should be reconciled to source reports, significant balances should be confirmed or inspected, and unusual allowances should be challenged. Inventory counts should be compared with perpetual records, count sheets, and subsequent usage, with differences evaluated for cutoff or shrinkage rather than automatically written off.

Intercompany testing must prove that bilateral balances and transaction amounts match, and that elimination entries are complete. A difference between one entity’s receivable and another entity’s payable can move a liability out of one ledger while leaving the corresponding asset in the other. Consolidation testing should trace trial balances to approved source systems, verify that eliminations are current, check that closing entries are posted across every entity, and reconcile reporting packages to consolidated totals. At the financial-statement level, cross-footing, comparative analysis, cash-flow consistency, note cross-references, and changes in estimates should be tested. These procedures can reveal a problem that balance-level reconciliation alone will not expose.

Common Mistakes That Make Testing Look Better Than It Is

A frequent mistake is asking the preparer to supply both the record and the proof of review. That can produce a completed checkbox without independent challenge. Reviewers should inspect source documents and rerun calculations, not just confirm that a spreadsheet contains the expected number. Another error is sampling only routine journals because they are easy to retrieve, leaving manual top-side entries, systems administrators, and unusual payments underrepresented. If automation extracts a “completed” status from the workflow, someone must also test who completed it, whether the action was authorized, and whether the supporting calculation was accurate.

Companies also misuse materiality. A threshold intended to identify material individual adjustments may miss multiple small errors, related transactions, and estimates with correlated uncertainty. Conversely, a zero-tolerance standard creates so many immaterial exceptions that reviewers become conditioned to accept them. Thresholds should reflect account risk, aggregation, and the purpose of the test, with common starting points for documentation ages often around 10, 30, or 60 business days. Those are governance examples, not accounting rules; the appropriate age depends on the account and normal settlement cycle. A small company should not copy the policy of a multinational without considering its staffing and information systems.

Copying prior-month working papers is another serious weakness. The review may be signed and dated correctly, yet it contains the prior period’s bank statement, stale exception analysis, or obsolete entry support. Working papers should carry enough information to identify what was tested in the current period, and a roll-forward mechanism should be used when a document changes little. Finally, teams often close exceptions without measuring whether the same issue returns. A control is not effectively remediated merely because an adjustment was posted; management should show that the source process changed, retest the change, and monitor subsequent periods for recurrence.

When to Escalate, Correct, or Seek Independent Help

Escalation should depend on potential impact, control weakness, management override, suspected fraud, regulatory exposure, and recurrence. A difference between two subsidiaries should be resolved immediately even if the consolidated elimination appears harmless, because the entity statements may be used for tax, borrowing, covenant, or local reporting. Unresolved bank reconciling items should be aged at every close; an item outstanding beyond the documented threshold normally requires an owner and written disposition. Journal entries posted outside normal access controls, duplicate payroll payments, unexplained related-party balances, or unsupported changes to master data should be elevated rather than handled as routine close noise.

Management should first expand the review to determine the full population affected. Then correct the current period, identify the proper accounting period, update affected statements or disclosures, and assess whether prior filings or tax returns require amendment or disclosure. A deadline-driven decision should not be made without knowing whether the error is material, intentional, or immaterial. If fraud, legal privilege, potential securities-law issues, or significant disagreement with management is present, counsel and qualified forensic or audit professionals may be needed. External auditors do not own management’s investigation, but their involvement can help establish independence and the audit response.

Corrective action should include a durable cause, accountable owner, due date, and evidence of retesting. A common sequence is contained by month-end day 2, reconciled and reported by day 5, investigated by day 10, and remediated within 30 to 90 days, with stronger deadlines for cash, fraud, and material issues. Those intervals are operational targets rather than universal requirements. The decisive question is whether the discrepancy could change decisions, conceal a loss, or indicate that management cannot rely on the same control next month. If the answer is yes, waiting for a later quarterly audit is the wrong response.

Cost, Staffing, and the Business Case

Month-end close testing has four main cost categories: staff time, accounting software, control consultants or external specialists, and the cost of delayed or corrected reporting. A spreadsheet-based program may have little direct licensing cost but can become expensive through rework, key-person dependency, and missed exceptions. Commercial close-management products span roughly tens to hundreds of dollars per user per month for basic workflow functionality, while more advanced reconciliation, controls, and financial-close platforms can cost thousands to tens of thousands of dollars annually for a smaller organization and considerably more at enterprise scale. Pricing varies by entities, modules, implementation, data hosting, and implementation services, so published ranges should be treated as market planning figures rather than quotes.

The business case should use quantified baseline data, not promotional estimates. Record the current close cycle, full-time-equivalent effort, number and age of open reconciliations, manual journal rate, unsupported adjustments, repeat exceptions, and average correction time. A hypothetical team that spends 300 staff hours monthly on close review, uses four analysts, and loads each with a fully loaded cost of $75 per hour has a direct labor baseline of $22,500 per month, or $270,000 annually. If better testing and workflow design reduce review time by 15%, the theoretical labor saving is $40,500 annually, before software and implementation costs. Any claimed return should therefore be compared with actual license, integration, training, and exception-management expenses.

Automation rarely eliminates judgment. It can gather populations, match data, flag anomalies, and preserve evidence, but finance professionals must determine whether a flagged item is valid and whether the underlying accounting treatment is appropriate. The strongest return usually comes from automating high-volume, clearly defined tests while reserving human review for estimates, unusual transactions, override risk, and disputed conclusions. A small organization may obtain adequate coverage through disciplined templates and independent review, while a multi-entity group may gain more from automated lineage and continuous controls monitoring. The right standard is reliable evidence at a sustainable cost, not the largest number of dashboard charts.

What Good Month-End Close Evidence Should Demonstrate

A mature testing process produces evidence that supports both completeness and accuracy. Completeness means that all required entities, accounts, journals, approvals, and consolidation inputs were included. Accuracy means that source records were correctly recorded, calculated, classified, cut off, and summarized. Timeliness means that exceptions were identified and resolved before the applicable reporting or decision deadline. Consistency means that the same control operated in the same manner across periods or that justified differences are documented. Independence means that the reviewer had enough access and authority to challenge the preparer rather than merely observing the work.

Useful performance indicators include the percentage of material accounts reconciled by deadline, the number and value of bank reconciling items by age, the percentage of manual journals with complete support, the count of repeat control exceptions, the median time to correct an error, and the value of proposed or posted adjustments. Reporting only a close score of 95% is ambiguous unless the formula and severity weighting are disclosed. A missed material fraud indicator should not be offset by hundreds of completed low-risk tasks. Metrics should therefore distinguish prevention, detection, correction, and recurrence, and should be traceable to source evidence.

Ultimately, month-end close testing is financial auditing applied to the management process before an external opinion is issued. It gives decision-makers confidence that the numbers are supported and that weaknesses are visible. The most defensible approach combines independent review, clear populations, source-to-ledger tracing, risk-based thresholds, and a controlled remediation process. A business that does this is not claiming every figure is perfect; it is demonstrating why material discrepancies are unlikely to remain hidden. That distinction is both more credible and more useful to boards, auditors, lenders, investors, and finance teams themselves.