By CA Surekha Ahuja
A practical year-end framework for CFOs, finance teams, and finance professionals
Financial statements are not final merely because the trial balance agrees. They are final when the numbers, underlying evidence, estimates, and disclosures together present a consistent and supportable picture of the business.
Year-end finalisation is usually treated as a mechanical closing exercise:
Trial Balance → Adjustments → Schedules → Financial Statements → Approval
That sequence is necessary. It is not sufficient.
A trial balance can agree perfectly while revenue is recognised in the wrong period, a material expense goes unrecorded, an old receivable is no longer collectible, an asset is carried above its supportable value, profit rises while cash generation weakens, or the notes tell a different story than the underlying records.
The real purpose of finalisation isn't to close the books — it's to establish whether the financial statements are complete, supportable, analytically consistent, and appropriately disclosed. A stronger sequence looks like this:
Complete → Reconcile → Analyse → Evidence → Judge → Adjust → Disclose → Approve
Order matters here. Analysis built on incomplete records misleads. Adjustments made without adequate evidence create new errors. And correct accounting without complete disclosure still adds up to incomplete reporting.
The five questions that should drive the final review
Before signing off, the finance team should be able to answer five questions clearly:
| Question | What it tests |
|---|---|
| Is the reported income real and attributable to this year? | Revenue recognition and cut-off |
| Have all material costs and obligations been recognised? | Completeness of expenses and liabilities |
| Are the reported assets genuinely recoverable and supportable? | Balance-sheet quality |
| Does reported profit convert into cash over a reasonable period? | Earnings quality and liquidity |
| Do the numbers, notes, and disclosures tell one consistent story? | Reporting integrity |
These five questions do more work than reviewing every ledger balance in isolation — they shift the process from checking numbers to understanding the business behind the numbers.
1. Start with completeness and reconciliation
The first discipline is simple: don't run sophisticated analysis on incomplete or unreliable data. Before the analytical review begins, confirm the major accounting records are substantially up to date.
| Area | Key review |
|---|---|
| Bank and cash | Reconciliation and verification of significant reconciling items |
| Receivables | Confirmation, reconciliation, ageing, subsequent collections |
| Payables | Reconciliation and review for unrecorded liabilities |
| Inter-company balances | Confirmation and elimination of unexplained differences |
| Statutory balances | Reconciliation with underlying returns and records |
| Inventory | Record completeness, physical verification, ageing |
| Fixed assets | Agreement between asset register and general ledger |
| Borrowings | Reconciliation with lender statements and agreements |
| Suspense/miscellaneous balances | Clear explanation and resolution of material items |
Reconciliation establishes consistency. It does not, by itself, establish correctness — and that distinction matters for everything that follows.
2. Review movements, not just closing balances
A closing balance tells only part of the story. Start instead from the movement:
Opening Balance + Additions − Reductions ± Reclassification = Closing Balance
For every material movement, ask: What changed? Why? Does the explanation make commercial sense? What evidence supports it? Does it affect recognition, measurement, or disclosure?
Track this for revenue, receivables, inventory, major expenses, provisions, and borrowings. The most useful question during finalisation is rarely "What is the balance?" — it's "Why did it move?"
3. Test the economic reality of revenue
Revenue is usually the starting point of analysis because errors here ripple through profit, receivables, taxes, and disclosures. The real question isn't whether the sales ledger agrees with the general ledger — it's did the income genuinely belong to this financial year?
Watch for: unusual monthly spikes (what commercial event explains them?), large year-end invoices (was delivery or service actually completed?), new significant customers, post-year-end credit notes, receivables growing faster than revenue, and high customer concentration.
The trail should hold together: Revenue → Delivery/Service → Receivable → Collection. Not every sale needs to result in immediate collection — but every sale should have an understandable economic trail. Where the trail breaks, dig deeper.
4. Test whether expenses and liabilities are complete
The most common year-end risk usually isn't a misrecorded expense — it's one that was never recorded at all. That means the review has to look past the reporting date: invoices received late, services consumed but not yet invoiced, employee obligations, interest and finance costs, professional and legal costs, recurring operating costs, contractual commitments, and material claims or disputes.
The governing question: if the obligation or consumption relates to this year, has it been recognised?
Provisions need their own discipline. For each material provision, document the nature of the obligation, the basis of the estimate, the key assumptions, the evidence available at year-end, and any subsequent developments relevant to those assumptions. A provision shouldn't survive purely out of habit — and it shouldn't be reversed purely because reversal flatters this year's profit. Consistency is not a substitute for reassessment.
5. Examine the quality of assets, not just their existence
For the balance sheet, the question isn't "does this balance exist in the ledger?" — it's what evidence supports its carrying value?
- Receivables: ageing, subsequent collections, disputes, customer financial position, concentration of exposure.
- Inventory: physical existence, slow-moving or obsolete stock, unusual build-up, the relationship between inventory and sales.
- Advances and other recoverables: flag balances that are old, unchanged for long periods, poorly documented, or hard to tie to a clear business purpose. Age doesn't make an unexplained balance safer — it may just mean the issue has survived several year-end closes undetected.
- Fixed assets and intangibles: major additions, disposals, capitalisation decisions, useful lives, depreciation, and any indicators affecting recoverability.
The sharpest test: would this carrying value still be supportable if reviewed for the first time today?
6. Read profit together with cash flow
Profit and cash flow answer different questions — a company can report strong earnings while liquidity quietly tightens. So trace the bridge from profit through receivables, inventory, payables and other liabilities, and non-cash items and provisions, down to operating cash flow.
| Pattern | Possible area for review |
|---|---|
| Profit ↑, operating cash flow ↓ | Earnings quality or working-capital pressure |
| Receivables ↑ sharply | Collection or revenue-recognition risk |
| Inventory ↑ while sales stay weak | Slow-moving or obsolete stock |
| Payables ↑ significantly | Liquidity pressure or delayed payments |
| Borrowings ↑ despite profits | Weak internal cash generation |
None of these patterns proves an error on its own. But every significant divergence deserves a credible explanation. Profitability and liquidity are related — they are not the same thing.
7. Review significant estimates as standalone items
Some of the most consequential numbers in the accounts — expected credit losses, impairment, provisions, useful lives, fair values, employee obligations, deferred tax assets — come from judgment, not invoices. A material estimate shouldn't live only inside a spreadsheet; it needs a documented trail: what's being estimated, what assumptions are used, what evidence supports them, what changed from last year, and how sensitive the outcome is to those assumptions.
A documented estimate isn't automatically a reasonable one — the assumptions still have to hold up commercially and evidentially.
8. Give related parties and unusual transactions extra scrutiny
Related-party transactions, inter-company balances, common counterparties, unusual financing arrangements, and large or non-routine year-end transactions all warrant review beyond ordinary ledger checks: who's the counterparty, what's the commercial purpose, are the terms documented, are balances reconciled, and are approvals and disclosures complete?
A useful gut-check: would this transaction look materially different if it had been entered into with an unrelated party? The answer doesn't determine the accounting treatment by itself, but it flags where closer scrutiny belongs.
9. Finalise disclosures with the same discipline as the numbers
One of the most avoidable weaknesses in year-end closing is treating disclosures as a drafting task done after the accounting is finished. Disclosures are part of financial reporting, not an appendix to it.
Follow the chain from ledger → supporting schedule → accounting analysis → note → financial statement, and for material balances check consistency of amounts, comparatives, cross-references, policy application, related disclosures, and explanations for significant movements. If a note explains a transaction differently than the underlying records do, either the accounting or the explanation needs a second look.
10. Review events after the reporting date before closing the file
Don't finalise without a structured look at what's happened since year-end: customer defaults, major credit notes, litigation developments, significant losses, refinancing, major contracts, acquisitions or disposals, and anything affecting liquidity or going concern.
The key question: does the subsequent event provide evidence about a condition that already existed at the reporting date, or does it relate to a new condition arising later? The answer can change recognition, measurement, or disclosure.
11. Run a red-flag matrix before sign-off
| Red flag | Question to investigate |
|---|---|
| Revenue rises sharply near year-end | Was the underlying obligation completed? |
| Receivables grow faster than revenue | Are balances genuinely recoverable? |
| Significant credit notes arise after year-end | Do they relate to year-end transactions? |
| Material expenses booked after year-end | Did they relate to the closed year? |
| Provisions are frequently reversed | Is the estimation process reliable? |
| "Exceptional" costs recur | Are they genuinely exceptional? |
| Inventory rises while sales stagnate | Impairment or obsolescence risk? |
| Old advances remain unchanged | Is recovery genuinely expected? |
| Profit rises but operating cash flow weakens | What explains the divergence? |
| Large period-end journals are posted | What's the commercial and accounting basis? |
| Round-sum adjustments are material | Is there adequate evidence? |
| Related-party movements shift significantly | Are approvals and disclosures complete? |
| Major estimates change | What evidence supports the revised assumptions? |
| Notes differ from schedules | Which is correct, and why? |
A red flag isn't proof of an error — it's a trigger for evidence-based investigation.
12. Where AI can strengthen finalisation — and where it can't
AI adds real value here, but only with its role clearly bounded. It's well suited to surfacing unusual trends, large or unusual journal entries, unexplained movements, ageing patterns, inconsistencies between schedules and notes, and unusual relationships between profit and cash flow.
The framework that keeps this useful: AI → Exception → Evidence → Professional judgment → Action, with four stages kept distinct — observation (what does the data show?), inference (what could explain it?), evidence (what supports or contradicts that explanation?), and conclusion (what action does this require?).
The biggest AI risk: it can generate a detailed, plausible-sounding explanation even when the information it had access to was incomplete. Every significant AI-assisted finding should be tested against three questions — what was actually reviewed, what relevant information was missing, and what independent evidence backs the conclusion. A plausible explanation is not evidence. And sensitive financial data still needs to be handled within proper confidentiality, access-control, and governance arrangements.
The final approval test
Before the statements move for approval, the team should be able to answer each of these clearly and with evidence:
| Final question | What it establishes |
|---|---|
| Does the reported profit make commercial sense? | Economic reality |
| Has all material income and expenditure been appropriately recognised? | Completeness and cut-off |
| Are major assets genuinely recoverable? | Balance-sheet quality |
| Does profit convert into cash over time? | Earnings quality |
| Are significant estimates supported by evidence? | Quality of judgment |
| Are related-party and unusual transactions properly addressed? | Transparency |
| Have significant subsequent events been reviewed? | Reporting completeness |
| Do the notes agree with the underlying records? | Disclosure integrity |
| Can every material unusual movement be explained? | Overall reliability |
If a material question can't be answered clearly and supported with evidence, the accounts aren't ready yet — no matter how well the trial balance ties out.
The bottom line
Finalising financial statements isn't just about proving that Assets = Liabilities + Equity. A set of accounts can balance perfectly and still need significant work. The real test is whether the statements are reliable, complete, supportable, consistent, and appropriately disclosed.
AI can help by processing more information and surfacing exceptions faster than a manual review alone. But the responsibility stays human: numbers identify the issue, analysis asks the question, evidence supports the answer, and professional judgment reaches the conclusion. That's the actual discipline of finalisation.
