Three-Way Trust Reconciliation, Explained Properly

Almost every firm that has ever been asked for a trust reconciliation can produce something. The problem is that what most firms produce is a two-way reconciliation wearing a three-way label. They compare the bank statement to the firm's own running total of trust cash, find that the two agree, and file the printout. That comparison is genuinely useful, but it only proves that the firm knows how much money is sitting in the account. It proves nothing at all about whose money it is.

The third leg is what makes the exercise a compliance control rather than a bookkeeping habit. A three-way reconciliation compares the adjusted bank balance, the trust control account balance in the firm's books, and the arithmetic sum of every individual client sub-ledger, all as of the same moment. If the first two agree but the third does not, the firm is holding the right amount of money and has lost track of which client owns which part of it. That is the failure mode that turns into a bar complaint, because a firm in that position cannot honestly answer the only question a regulator really cares about, which is whether every client could be paid out in full today.

This post walks through what each of the three balances actually is, why they must agree to the cent rather than approximately, what cadence regulators in the common-law markets expect, and the specific, recurring mistakes that break the match. The rules differ meaningfully between US states, between Canadian provinces, between Australian states and against the position in England and Wales, so treat the mechanics here as universal and the timing and record-keeping specifics as something you confirm against your own regulator.

The first balance: the bank, adjusted rather than raw

The first number is not the figure printed at the bottom of the bank statement. It is that figure adjusted for items the bank has not yet processed but which the firm has already recorded. Outstanding cheques written to a client or a court but not yet presented get subtracted. Deposits recorded in the books on the last day of the period but not credited by the bank until the next business day get added. What you end up with is the amount the bank would show if every transaction the firm has legitimately recorded had already cleared.

That adjustment list is not a nuisance, it is evidence. A properly prepared reconciliation names each outstanding item with its date, its amount and the matter it belongs to, so that anyone reading it can see exactly why the bank's number differs from the firm's. A reconciliation that closes the gap with a single unlabelled adjusting figure is not a reconciliation, it is a plug. Regulators and forensic accountants have seen that plug many times and they open it first, because it is where genuine shortfalls hide.

The second balance: the trust control account in your own books

The control account is the firm's single running total of all client money held in a given trust bank account. Every deposit increases it, every disbursement decreases it, and it should never be edited by hand. In practice this is where a lot of small firms go wrong, because a control total maintained in a spreadsheet can be typed over, dragged, sorted incorrectly or have a row deleted without leaving any trace at all. The number still looks like a control total afterwards. It just is not one any more.

The control account exists so there is a second, independent path to the same figure. If it is derived from the sub-ledgers by formula, it is not independent and the third leg of the reconciliation collapses into a tautology. In Casely the trust ledger is maintained as posted entries rather than as an editable balance field, and any disbursement that would push a matter past its actual trust balance is blocked at the database transaction level rather than surfaced as a warning dialog the user can click through. That distinction matters here because a control total can only be trusted if the entries feeding it could not have been written in an invalid state to begin with.

The third balance: the sum of every client sub-ledger

Each matter holding client money needs its own ledger showing that client's deposits, that client's disbursements and that client's running balance, in date order. Add every one of those closing balances together and you should land on exactly the control account figure and exactly the adjusted bank figure. This is the leg that answers the ownership question, and it is the leg most commonly skipped, because on a busy month-end it feels redundant. It is not redundant. It is the entire point.

The sum must include matters with a zero balance and, crucially, must not quietly exclude any matter with a negative balance. Per-matter isolation is what makes the arithmetic honest, because when each matter carries its own ledger rather than a filtered view of one shared list, a deficit on one file cannot be masked by a surplus on another. A firm that keeps client money in a single undifferentiated pool and reports one total has, functionally, no third balance to reconcile.

Why "close enough" is not a standard that exists here

In most areas of firm finance a small unexplained variance is an annoyance you carry forward. In trust accounting it is a red flag by definition, because every cent in that account belongs to an identifiable client and an unexplained difference means at least one client's recorded entitlement is wrong. A twelve dollar discrepancy is not twelve dollars of risk. It is a signal that the process which produced it is capable of producing a much larger error, and that no one has yet found the cause.

There is also a hard asymmetry that catches firms out. A surplus is not the safe direction. Money sitting in trust that belongs to nobody, or that belongs to the firm, is its own violation in most jurisdictions, and leaving it there to "cover" future differences is exactly the behaviour rules against commingling exist to prevent. The correct response to an unexplained surplus is to investigate it, document what you find and follow your regulator's process for unclaimed or unidentified funds, which in many jurisdictions eventually means remitting it to a designated body rather than absorbing it.

The cadence regulators actually expect

The common pattern across common-law markets is a reconciliation performed at least monthly, prepared promptly after the period closes, signed off by someone with responsibility, and retained for a defined number of years. In the United States the requirement sits with each state bar and the specifics vary considerably in both frequency and record retention, so a firm licensed in more than one state should not assume the strictest rule it knows applies everywhere. In England and Wales the SRA Accounts Rules have long required client account reconciliations at intervals of no more than five weeks, which is a subtly different obligation from "monthly" and catches firms that drift to a fixed calendar date.

Canadian law societies generally require a monthly trust comparison prepared within a set window after month end, with the details set provincially, and Australian trust rules are similarly set at state and territory level under the uniform law framework where it applies. The practical takeaway is not the specific interval, it is that the interval is short, the deadline for completing the work is real, and the record is expected to exist contemporaneously rather than be reconstructed later. Confirm the exact frequency, deadline, sign-off and retention period with your own regulator, because getting this wrong is a self-reporting problem before it is anything else.

Mistake one: timing differences treated as errors, and errors treated as timing

The single most common reason a reconciliation appears to fail is a genuine timing difference, a cheque issued on the twenty-eighth that clears on the third. The danger is not the timing difference itself, it is the habit it creates. Once a bookkeeper is used to seeing a gap and attributing it to timing, a real error that happens to be similar in size gets waved through with the same explanation. The discipline that prevents this is refusing to net anything: every outstanding item is listed individually, by date, amount and matter, and the list must resolve item by item in the following period.

The reverse failure is just as damaging. A stale outstanding cheque that has sat unpresented for six months is no longer a timing difference, it is an unresolved liability, and carrying it forward silently makes every subsequent reconciliation slightly less true. Any item that ages past a threshold you set, and many regulators effectively set one for you through stale-dated cheque rules, needs to be investigated and either reissued or written back with a documented reason rather than rolled forward again.

Mistake two: bank charges, interest and reversals posted with no client attached

Bank-originated entries are the classic breaker of the third leg. A monthly account fee, an interest credit, a returned item charge or a wire fee hits the bank statement and therefore the bank balance, and if someone posts it to the control account to make the first two numbers agree, there is now no client sub-ledger carrying it and the sum of sub-ledgers falls out of line. The fix is never to force the control total. It is to determine whose money, if anyone's, that entry properly affects, and to handle firm-borne charges through the firm's own operating account in the way your jurisdiction requires.

Interest deserves particular care because the treatment genuinely differs by jurisdiction and by account type. Pooled interest-bearing arrangements in several markets direct interest to a legal foundation or public-benefit body rather than to the firm or the client, while separately designated accounts for a single client usually credit that client. Getting this wrong does not just break the reconciliation, it misallocates money that was never the firm's to allocate, so confirm the rule for your specific account type locally rather than copying what a firm in another jurisdiction does.

Mistake three: a negative client balance hidden inside a healthy total

This is the failure that three-way reconciliation exists to catch, and it is quietly common. A disbursement is made on one matter for more than that matter holds, the overall account still has plenty of cash because other clients' funds are sitting there, and the bank balance and control total continue to agree perfectly. Only the sub-ledger sum reveals the problem, and only if negative balances are shown rather than suppressed. In substance, one client's money has been used to fund another client's transaction, which is among the most serious findings a trust audit can produce.

The structural answer is to make the state unreachable rather than detectable after the fact. When each matter carries an isolated ledger and any disbursement exceeding that matter's actual balance is rejected inside the same database transaction that would have written it, there is no window in which a negative balance exists to be found later. That is a meaningfully different guarantee from software that records the entry and flags it, because a flag depends on someone reading it, and month end is exactly when nobody does.

Mistake four: corrections made by editing history instead of reversing it

When a bookkeeper finds a wrong entry, the instinct is to fix the entry. In trust accounting that instinct is wrong. Editing or deleting a posted entry changes what the ledger says happened, which means the reconciliation you signed last month no longer reproduces from the underlying data, and an auditor comparing the two will reasonably ask what else changed. The correct treatment is to reverse the original with a dated correcting entry that references it, leaving both visible.

Casely handles this by voiding rather than deleting, so a corrected entry stays in the record with its correction attached and the audit trail reads as a sequence of events rather than a tidied-up final state. Documents supporting the reconciliation carry a comment field recording what changed and why, which is what turns a stack of PDFs into an explanation an examiner can follow without interviewing your bookkeeper. Every stored document is encrypted with AES-256 under a per-firm key, which matters because trust reconciliations and the bank statements behind them are among the most sensitive financial records a firm holds.

Mistake five: earned fees left sitting in trust

Money that the firm has genuinely earned and is entitled to draw is firm money, and most jurisdictions expect it to be moved out of trust promptly once the client has been billed and any required notice period has passed. Leaving it there feels conservative and is not. It inflates every one of the three balances with funds that no longer belong to a client, it complicates the sub-ledger arithmetic, and in several jurisdictions it is itself a commingling problem. The reconciliation will still tie, which is precisely why this one survives so long undetected.

The practical cause is almost always billing latency rather than intent. Fees cannot be transferred until they are billed, so a firm that invoices sporadically accumulates earned money in trust by default. Turning unbilled time into an itemised draft in one click removes the excuse, because the constraint stops being "we have not got round to billing" and becomes a genuine decision. Hourly, flat-fee, contingency and blended arrangements each have their own timing for when money becomes the firm's, so the billing model and the trust drawdown rule need to be understood together rather than separately.

What the reconciliation record itself should contain

A reconciliation is not the summary page. The defensible record is the summary plus the three supporting listings that produce it: the bank statement with the outstanding items schedule, the control account activity for the period, and the full client ledger listing with individual balances and a total. If any of those three is missing, nobody can verify the summary independently, which means the summary is an assertion rather than evidence. Keep them together, keep them in the same place, and keep them for as long as your regulator specifies.

The record also needs a date of preparation and a named person who reviewed it, and in many jurisdictions that reviewer is expected to be someone other than the person who maintains the ledger day to day. In a two-person office that separation is imperfect, and the honest response is to document who did what rather than to pretend to a segregation of duties that does not exist. An examiner is far more forgiving of a small firm that clearly describes its limitations than of one that implies a control it never operated.

Why the sign-off is not a formality

The sign-off exists because the reconciliation is where a shortfall becomes knowable. In most jurisdictions the obligation to correct a client account shortfall promptly, and in some cases to report it, is triggered by discovery, and the reconciliation is the discovery mechanism. A reviewer who signs without actually testing the outstanding items list or spot-checking a few sub-ledger balances has not just skipped a step, they have converted an unknown problem into a documented one that the firm then failed to act on.

That is also why reconciliations should never be batched. Three months completed in one sitting in April are not three monthly reconciliations, they are one annual scramble with three dates on it, and the gap between when an error occurred and when it was found is exactly what regulators look at when deciding how seriously to treat it. A firm that reconciles on schedule and finds a genuine error is in a completely different position from a firm that finds the same error six months late, even though the error is identical.

Making the three balances agree without it eating a week

The reason three-way reconciliation feels heavy is almost never the reconciliation itself. It is the reconstruction work in front of it, chasing which matter a deposit belonged to, working out whether a disbursement was ever properly recorded, and rebuilding sub-ledgers that were never really maintained as sub-ledgers. When the underlying system posts every trust movement against a specific matter at the time it happens, refuses entries that would create an impossible state, and keeps corrections visible instead of overwriting them, the month-end exercise becomes a comparison rather than an investigation. That is the difference between an hour and a week.

If your current process depends on a spreadsheet that anyone can edit, the honest assessment is that you have a report rather than a control, and the fix is structural rather than procedural. It is worth looking at how trust accounting software for law firms enforces per-matter isolation and blocks overdrawing at the data layer, and worth looking separately at legal billing software if earned fees tend to linger in trust because invoicing lags the work. Those two problems look unrelated on an org chart and are the same problem on a reconciliation.

None of this removes the need to know your own rules. Frequency, deadlines, who may sign, how long records are kept, how interest is treated and what happens to unidentified funds all vary by state, province and territory, and they change. Get the mechanics right first, because they are the same everywhere, then confirm the specifics with your regulator and build them into the calendar rather than into somebody's memory. Casely starts at $0 on the Free plan and is cloud-native with no local install, so testing whether your trust records would actually tie three ways is a shorter exercise than most firms assume.