The particular difficulty of inherited work
Closing a year you kept yourself is largely a processing problem. You know why each decision was made, which balances are soft, and where the bodies are buried, because you buried them. The work is bounded, and the unknowns are ones you chose.
Inherited work is a different activity wearing the same name. The processing is identical; the difficulty is that every figure you did not produce is an assertion by someone whose reasoning you cannot see. An opening balance is not a fact — it is a conclusion, reached by a person you may never speak to, using information you may not have.
That reframing changes the order of the work. Before touching a transaction, establish which inherited figures you are going to test, which you are going to accept, and — critically — write down that you accepted them. The commercial and professional risk in this kind of engagement is almost never in the processing. It is in the silent adoption of someone else's conclusions.
Establishing the opening position
The opening position is everything the books assert as at the moment you took over. It is the foundation the whole year rests on, and if it is wrong the year cannot be right no matter how carefully you process it.
Some of it is testable against outside evidence. Cash can be proved against bank statements. Loans can be proved against lender statements. Some balances can be confirmed with third parties. Everything else — accruals, provisions, work in progress, stock, the valuation of anything — rests on judgement, and judgement cannot be verified, only re-performed or accepted.
So split the opening trial balance into three: what you will verify, what you will accept, and what you will flag as unable to verify. The third category is not a failure. It is the honest description of most inherited engagements, and stating it plainly is what protects both you and the client.
The line that matters
What to ask for on day one
Ask immediately, because the predecessor's willingness to respond declines steeply with time, and the client's memory of what existed declines with it.
The trial balance at the handover date. The prior year's finalised accounts. The last completed reconciliation for every bank account, with its date. The complete statement series for the period you are responsible for. The list of open reconciling items, with the date each arose. An explanation of any suspense, holding or clearing account. And the list of bank accounts, cards and payment platforms that exist — which is more often incomplete than any other item on this list.
If nothing arrives, proceed anyway. Statements come from the bank, filed accounts are frequently public, and clients hold more documentation than they believe. Rebuild from sources you can obtain independently, and record what could not be obtained and from whom — that record is worth writing at the time and impossible to reconstruct later.
| Ask for | Why it matters | If you cannot get it |
|---|---|---|
| Trial balance at handover | The opening position you are adopting | Rebuild from the prior year's accounts |
| Prior year finalised accounts | The last agreed position | Often publicly filed and obtainable |
| Last completed reconciliations | Tells you the date cash was last proved | Reconcile back yourself until it agrees |
| Full statement series | The independent record of the year | Request from the bank directly |
| Open reconciling items with dates | Distinguishes timing from stale error | Derive from the last reconciliation |
| Explanation of suspense accounts | The unanswered questions you inherit | Work the individual entries, not the total |
| List of every account and card | Scope — routinely incomplete | Look for transfers with no counterparty |
Rebuild cash before anything else
Cash first, always. Not because it is the largest number or the most interesting, but because it is the only account in the entire ledger with an independent third-party record that can be obtained without anyone's cooperation.
Every other balance ultimately rests on documents the client holds, or on judgements someone made. The bank statement rests on neither. That makes cash the one place you can convert an inherited assertion into a verified fact, and a single anchored fact is worth a great deal when everything else is uncertain.
It also has the widest reach. Cash touches revenue, costs, payroll, tax payments, borrowings and drawings. A fully rebuilt and reconciled bank position constrains all of those, so errors elsewhere become visible as things that cannot be true given what the bank shows.
And it is the part that automates best. A year of statements is a volume problem, not a judgement problem — which means it is exactly the part of the engagement you should not be doing by hand while the judgement work waits.
Assembling the statement estate
Gather every statement for the year, for every account, including the ones nobody mentioned. Old cards, a savings account, a payment platform holding a balance, an account opened for one project — handover lists are routinely incomplete, and an account nobody mentioned is an account nobody reconciled.
Then run two checks before any of it reaches the ledger, because they answer different questions. Within each document: opening balance plus every transaction should equal the printed closing balance. Between documents: each closing balance should equal the next one's opening balance.
The first proves each statement is whole. The second proves the series is unbroken. A handover folder can pass either one while failing badly on the other, and both failures are common precisely because the folder was assembled retrospectively by someone under time pressure. The series test is covered in detail in the missing month problem.
Convert rather than key. Handover periods combine the highest volume with the lowest familiarity, which is precisely the conditions under which transposition errors are introduced — and those errors are then indistinguishable from the predecessor's mistakes, which is a genuinely unpleasant place to end up.
When the ledger and the bank disagree at the start
Frequently the opening bank balance in the ledger does not match the bank. The temptation is to adjust it to reality and proceed — after all, the bank is right. Resist it, at least until you know when the divergence began.
Walk backwards. Compare the ledger to the statements for the prior period, then the one before, until you reach a point where they agree. That point is where the error was introduced, and knowing it tells you which year is actually misstated — which may not be the year you are closing.
The distinction matters. An adjustment posted today for an error made two years ago charges the current year with something that belongs elsewhere, distorting the results you are about to report and the comparatives alongside them. A constant, unchanging difference is the classic signature of exactly this problem, and its diagnosis is covered in why reconciliations fail.
If the origin genuinely cannot be found, correct it in the current year with a written explanation of what was investigated and where you stopped. A documented decision is defensible. A silent adjustment is not, and it is the one that gets asked about.
Reconciling items you did not create
Handover reconciliations tend to carry a tail of open items, and the useful question is not what they are but how old they are. An uncleared cheque from last month is a timing difference. An uncleared cheque from three years ago is not — it is an entry describing something that never happened.
Age every open item and treat the old ones as errors rather than as pending. A payment that never cleared means the obligation was never settled, which may mean the liability still exists, or that the entry should never have been made. Both are findings. Neither is resolved by carrying the item into another year.
Deposits in transit that never landed deserve more urgency, because money that was recorded as received and never arrived is a different class of problem entirely. It may be a recording error; it may be that the money genuinely did not arrive. Establish which before closing anything.
Clearing this tail is one of the most visible improvements you can make, and one of the few things in an inherited engagement that is unambiguously better afterwards.
The suspense account you inherited
A suspense account with a balance is a list of unanswered questions that has been added up. The total is arithmetically real and analytically meaningless — it is the residue of several unrelated problems that happen to have been parked in the same place.
So work the entries, never the balance. Each one is an individual question: what was this, why could it not be classified, and what is the answer now. Some resolve immediately with the bank statement in front of you, because the underlying transaction is obvious once the description is visible.
What genuinely cannot be resolved should be written off deliberately, with an explanation, and with the client informed. Carrying an unexplained balance into another year is a decision to preserve a known problem, and it will be inherited again by whoever comes after you — usually with less context than you have now.
A useful tell
Decisions you disagree with
You will find categorisations you would not have chosen. The discipline is to distinguish decisions that are wrong from decisions that are merely different, because reworking the second kind consumes time that the current year has a stronger claim on.
Restate where it changes something: tax treatment, a covenant or ratio, a reported figure, or comparability between the year you are closing and the one before it. That last one matters more than people expect — accounts where the same cost sits in different places in consecutive years mislead every reader, even when both placements are individually defensible.
Leave alone where the predecessor made a reasonable choice you would have made differently. Perfect consistency with your own preferences is not a deliverable, and pursuing it in inherited books is one of the more reliable ways to overrun an engagement without improving the accounts.
Where you do restate, note it. A reader comparing two years should be able to see that a presentation changed and why, rather than concluding that the business did.
Positions that have already been filed
Sometimes the work reveals that a previously submitted return rested on figures you now believe are wrong. This is the point at which the engagement stops being a bookkeeping exercise and becomes a professional judgement with rules attached.
Those rules vary by jurisdiction and by professional body, and this is not the place to summarise them — follow your own body's guidance and the local requirements, and take advice if the position is significant. What is universal is that the client needs to know, in writing, and that the decision about what to do is theirs to take with your advice.
What is not acceptable anywhere is quietly rebuilding the comparatives so the discrepancy disappears. That removes the evidence of a question that genuinely needs an answer, and it converts someone else's error into your own act.
Practically: quantify the difference, document what you found and how, tell the client promptly, and record the decision that was reached. Then close the year you were engaged to close.
Documents that simply do not exist
Some supporting documents will be gone. Receipts nobody kept, invoices from a supplier that closed, a period where the filing system was somebody's inbox. This is normal in inherited work and it is not, by itself, a reason to stop.
The bank statement remains evidence that money moved, and it is independent evidence at that. It does not establish what was bought, which is a real limitation with tax consequences in most jurisdictions, but it does establish the movement and its counterparty — often enough to support a reasonable treatment.
Record what is missing rather than working around it silently. A schedule of unsupported items with amounts is a professional deliverable: it tells the client where their exposure sits, gives them the chance to find documents you could not, and makes the boundary of your work explicit.
And fix it forward. Most missing-document problems are collection problems, and the fix is a habit rather than a system — capturing documents when they arrive rather than assembling them at year end.
Verify, accept, or flag
A working default for the opening trial balance. Adjust for materiality and for what the engagement is actually paying for, but decide explicitly rather than by default — and record the decision, because the record is the deliverable that protects you.
| Inherited balance | Default treatment | Why |
|---|---|---|
| Cash and bank | Verify | Independent third-party evidence is obtainable |
| Loans and finance | Verify | Lender statements confirm it directly |
| Tax balances | Verify where statements exist | Authority records are usually obtainable |
| Receivables | Verify by sample and ageing | Old items are often uncollectible or already settled |
| Payables | Verify by sample | Understatement is the common direction |
| Stock and work in progress | Accept, flag prominently | Cannot be re-performed after the date |
| Accruals and provisions | Accept, flag | Judgement made with information you do not have |
| Fixed assets | Accept, review additions | Historic cost is rarely worth reconstructing |
| Suspense and clearing | Resolve entry by entry | Not a balance — a list of open questions |
A worked example
A company arrives in September needing a year closed to December. The previous bookkeeper left in April, the client says the books are "a mess", and the last completed reconciliation is dated the previous November.
Cash first. Statements are gathered for all three accounts, including a card nobody mentioned that turns up because transfers appear with no counterparty. The series test shows a gap in July for the main account — the download window had closed — and an overlap in the card statements, where the same period was pulled twice.
The July statement is requested from the bank and arrives in a week. Continuity then holds across all three accounts, and each document passes its internal identity. Cash is now a verified fact rather than an inherited claim, and it took days rather than weeks because the volume work was converted rather than keyed.
Reconciliation from November forward reveals the opening ledger balance was 1,240 below the bank. Walking back finds the origin two periods earlier: a duplicated import, half-corrected. It is corrected at source, so the current year is not charged with a prior year's error.
Suspense holds 3,880 across nine entries. Six resolve immediately once the bank descriptions are visible. Three do not, total 410, and are written off with a note. Opening stock and accruals are accepted and flagged in the engagement letter as unverified. The year closes, and the boundary of what was checked is written down.
| Stage | Finding | Action |
|---|---|---|
| Statement gathering | An account nobody mentioned | Added to scope |
| Series continuity | July missing; card periods overlapping | Requested July; deduplicated the card |
| Document identity | All statements whole once complete | Cash accepted as verified |
| Opening bank balance | 1,240 below bank, origin two periods back | Corrected at source, not in the current year |
| Suspense | 3,880 across nine entries | Six resolved, three written off with a note |
| Stock and accruals | Cannot be re-performed | Accepted, flagged in the engagement letter |
Scoping it in writing
Inherited engagements overrun more than any other kind, and the cause is almost always scope rather than speed. The processing is predictable; the investigation is not, and investigation is most of the work.
So scope it as two things. The closing of the year is a defined piece of work with a price. Investigating what preceded it is a separate activity with its own budget and its own stopping rule — for instance, verify cash and loans, accept the rest, and stop walking backwards after two periods without agreement.
State the stopping rule explicitly, including in the deliverable. "We traced the bank difference back two years and could not identify its origin; it was corrected in the current year" is a complete, professional answer. Silence about the same facts is what creates a problem later.
Leaving it better than you found it
Every difficulty in this article was created by a predecessor who documented nothing — not through negligence, usually, but because the reasoning was obvious to the person holding it and was never written down.
So leave four things. A clear opening position with a note on what was verified and what was accepted. Reconciliations completed to a stated date for every account. A short note on every judgement that is not evident from the ledger. And no open items you did not explain.
That set takes an hour at the end of an engagement and removes most of the cost of the next handover. It also happens to be exactly what makes your own file defensible if anyone asks about the year afterwards, which is a reasonable second motive.
| Leave behind | Cost to you | Value to the next person |
|---|---|---|
| Opening position with verified/accepted split | Twenty minutes | Removes the largest unknown entirely |
| Reconciliations completed to a date | Already done | Fixes the point cash was last proved |
| Notes on non-obvious judgements | Ten minutes | Prevents needless rework of sound decisions |
| No unexplained open items | Varies | Stops errors being inherited a second time |
| A list of every account and card | Five minutes | Prevents the account nobody mentioned |
Key takeaways
Inherited figures are assertions, not facts. The work is not to accept or reject them wholesale, but to decide deliberately which you will verify, which you will accept, and which you will flag — and to write that decision down, because silent adoption is where the real exposure sits.
Start with cash. It is the only account with independent third-party evidence you can obtain without anyone's cooperation, it constrains almost everything else, and it is the part that automates well. Verify the statement series is unbroken and each document whole before any of it reaches the ledger.
Resolve rather than carry. Stale reconciling items and inherited suspense balances are known problems, and moving them into another year is a decision to preserve them. Then leave a clear opening position behind you — the thing you wish you had been given.
Frequently asked questions
Rebuild a year of cash in an afternoon
Convert a whole year of statements in one run, merged into a single sheet with a source column, with every document checked against its own arithmetic as it goes. The volume work stops being the reason the investigation never starts.