Law Firm Accounting: The Three Sets of Books and What Each One Is For

Almost everything written about law firm accounting is written by someone who wants to sell the firm a service or a subscription, and it all starts in the same place: open a trust account, open an operating account, do not mix them. That advice is correct and it is nowhere near sufficient. A law firm does not keep one set of books. It keeps three, and they answer to three different authorities. The conduct rules specify, item by item, the records a lawyer must keep for other people’s money. The tax code specifies what the firm must be able to substantiate about its own. And then there is a third set that nobody requires at all, the numbers that tell an owner whether the firm is working, which is the set most small firms never build. Confusing the three is the reason a firm can be perfectly compliant, perfectly filed, and still unable to answer the simplest question about its own finances.

What law firm accounting actually is, and why it is three jobs rather than one

Start with the thing that makes a law firm different from a dental practice or a landscaping company of the same size. A law firm routinely holds money that belongs to someone else, in quantity, for long periods, and it is answerable for that money to a regulator with the power to end the owner’s career. Everything distinctive about legal accounting follows from that one fact.

The Oregon State Bar’s Professional Liability Fund, presenting a 2025 continuing legal education session on the subject, put the structure plainly. General accounting “tracks firm operations.” Trust accounting “ensures client/third party funds are safeguarded.” And then the sentence that most explanations skip: legal accounting is “a combination of general and trust accounting.” Not one or the other. Both, running in parallel, reconciling to each other, kept to different standards.

There is a third layer the compliance literature rarely names, because no authority requires it. Call it the management set: realization on the work you performed, the age of what you are owed, the cost of carrying a matter from engagement to payment, the share of revenue that reaches the owner. Nobody will discipline you for not having these numbers and nobody will audit them. They are simply the only numbers that answer the question an owner actually has, which is whether the firm is working.

The three sets of books a law firm keeps

Hold that structure for a moment, because it explains a pattern that confuses a lot of owners. A firm can pass a trust audit and still have no idea what a matter costs. A firm can file an immaculate return and still be unable to say which practice area carries the others. Compliance and insight are different products of different records, and buying software that does the first does not deliver the second.

It also explains who does what. The same Oregon presentation separates the roles in a way worth copying: a bookkeeper handles “daily financial management” and trust account tracking, a tax accountant handles “tax filing, high-level advisory,” and bookkeeping “must take place before any accounting.” Firms that hire a tax preparer and assume they have hired a bookkeeper discover the gap in March, when the preparer asks for records nobody has been keeping. Delegation does not move the duty, either. The same materials state the point plainly: “You are accountable” even where the work is delegated.

The books your conduct rules specify, line by line, and only for client money

Here is the asymmetry nobody points out. Open your jurisdiction’s trust records rule and you will find an itemised inventory. Washington’s RPC 1.15B is representative. It requires that “a lawyer must maintain current trust account records,” that they “may be in electronic or manual form and must be retained for at least seven years after the events they record,” and then lists nine categories at minimum.

Those nine run from a checkbook register for each trust account carrying, for every single movement, the client matter, the date, the check number, the payor or payee and the new balance; through individual client ledger records showing the same five facts plus the purpose; to copies of fee agreements, copies of accountings given to clients, copies of bills rendered, copies of the invoices supporting every disbursement, bank statements and deposit slips and cancelled checks, copies of all bank and client ledger reconciliations, and finally the portions of the client file “reasonably necessary for a complete understanding of the financial transactions.” A companion provision sets the reconciliation cadence: trust records “must be reconciled as often as bank statements are generated or at least quarterly,” matching the register to the bank and the register to the combined total of the client ledgers. Our treatment of lawyer trust accounting walks that reconciliation and those duties in full, and the IOLTA account mechanics sit alongside it.

The nine categories of trust record one adopted rule requires

Three features of that list repay a second look. It is a list of copies as much as of ledgers, which means a firm that keeps immaculate balances and discards the paperwork behind them has still not complied. It reaches the client file, selectively, for the documents that explain a transaction. And it is a minimum rather than a design, so a firm meeting it exactly has satisfied the regulator and learned nothing about itself. A Washington comment adds a practical gloss worth copying into any firm’s procedures: if the records are computerised, “a system of regular and frequent (preferably daily) back-up procedures is essential.”

Now look for the equivalent list governing the firm’s own money. There is not one. Your conduct rules care about the operating account mainly to the extent that firm money must not sit in the trust account and client money must not sit in yours. Everything else about your own books is left to tax law and to your own judgment, which is precisely why so many small firms have immaculate trust records and an operating ledger that is really just a bank feed with guesses attached.

Set the three sets side by side and the division of labour becomes obvious, as does the reason firms so often think they have done the work when they have only done a third of it.

The practical instruction that falls out of the table is short. Treat the first row as a floor set by someone else and meet it exactly. Treat the second as a question of substantiation rather than of filing. Treat the third as the only part you are actually designing, and design it on purpose rather than inheriting whatever categories a software default installed.

Cash or accrual, and the one thing a cash-basis profit and loss cannot tell you

Two methods, one choice, and most small firms make it without noticing they made it. Under the cash method, as IRS Publication 538 states, “you generally report income in the tax year you receive it” and “you deduct expenses in the tax year in which you pay the expenses.” Under an accrual method, “you generally report income in the year it is earned and deduct or capitalize expenses in the year incurred,” subject to the all-events test, which “is met when all events have occurred which fix your right to receive the income and you can determine the amount with reasonable accuracy.”

Law firms almost always file on cash, and almost always may. The general bar on the cash method reaches corporations other than S corporations, partnerships with a corporate partner, and tax shelters, and it is lifted for any entity that satisfies the gross receipts test of section 448(c). That ceiling is indexed for inflation and the IRS republishes it every autumn. For tax years beginning in 2026 a corporation or partnership meets the test “if the average annual gross receipts of such entity for the 3-taxable-year period ending with the taxable year which precedes such taxable year does not exceed $32,000,000.” The figure was $31,000,000 for 2025, $30,000,000 for 2024 and $29,000,000 for 2023.

The cash-method ceiling under section 448(c), by tax year

So the choice is open, and cash is the sensible default for most firms: you are taxed on what you collected rather than on what you billed, which matters a great deal in a business where some of what you bill is never collected. But the cash method buys that simplicity at a price, and the price is that your profit and loss statement is a record of bank activity rather than of work.

Consider a single matter. You open it in January, work it steadily through the year, and it settles in December. On the cash method your statements show twelve months of expense and one month of revenue. The matter looks ruinous for eleven months and miraculous in the twelfth, and neither picture is true. Now scale that to a practice where most matters behave this way and you have a profit and loss statement that cannot answer the question an owner most wants answered, which is whether the work itself pays.

The IRS’s own guide for examiners auditing attorneys records the workaround the profession actually uses. Attorneys, it says, “generally file tax returns using the cash basis, although some of the accounts for income earned, work in progress, and costs advanced will be kept on the accrual basis.” That is the practical shape of a good set of law firm books: file on cash, manage on something closer to accrual, and keep the bridge between them visible rather than improvised.

The same matter, the same year, two methods

Both lines finish at the same number. That is the point worth holding on to, because it is why the choice feels academic until the moment it does not. The year’s total is identical. What differs is everything you could have known during the year, which is the only period in which a decision was still available to you.

Advanced client costs, the entry that is not an expense

If there is one bookkeeping error that does more damage than any other in a small firm, this is it. You pay a filing fee, a court reporter, an expert, a records request. The money leaves your operating account. Your bookkeeper, reasonably enough, records an expense. It is not one.

The IRS states the position directly in its guide for examiners auditing attorneys. “Courts have determined that costs paid on behalf of a client are to be treated as in the nature of loans for tax purposes. They are not deductible by the attorney as a current cost of conducting business. The costs are those of the client and not the attorney since there is an expectation of reimbursement.” The conclusion the guide draws is unambiguous: “attorneys on the cash method of accounting may not take a current deduction for client expense advancements for which an attorney expects to be reimbursed.” A bad debt deduction becomes available in the year a particular advance is determined to be uncollectible, which is a different year and a different entry.

The same document tells you where the money belongs instead. Advanced costs “will normally appear in an asset account such as Unbilled Advanced Client Costs,” and in the firm’s systems they are “recorded into the system in a similar fashion to other accounts receivables.” That is the whole correction: an advance is a receivable, not a cost. It sits on the balance sheet until it is recovered or written off.

The categories involved are the ordinary ones. The guide quotes the Tax Court’s list from a 1969 decision: “travel expenses, costs of medical records, reports, interpreters’ fees, witness fees, deposition costs, filing fees, investigation costs, photographs, laboratory tests, and sheriff’s fees for service.” There is a genuine carve-out for genuinely general overhead. Attorneys on the cash method “are generally allowed a current deduction for client reimbursed costs which are allocated to normal operating expenses, (for example, secretarial costs or copying costs),” because these “would reasonably be incurred even if not charged to a particular client,” with any later reimbursement then treated as income under the tax benefit rule.

Where the same $40,000 of advanced costs lands

Look at the right-hand pair for a second longer than feels necessary. Under the wrong treatment, forty thousand dollars that clients owe the firm appears nowhere on the balance sheet at all. The firm has lent the money and has no record that it is owed. That is not primarily a tax problem. It is the reason firms chase costs they have already forgotten about, and the reason a departing partner or a prospective buyer cannot value what the firm is carrying.

There is a second reading of the same pair, and it is the one that matters at tax time rather than at management time. The expense treatment reports forty thousand dollars less profit than the firm made, which feels conservative and is not. It is a deduction the IRS guide says is not available, taken in a year in which the money was lent rather than spent, and it will be followed by a year in which the recovery arrives and the firm has nothing left to offset it against. The distortion runs in both directions and nets to nothing over the life of a matter, which is precisely why it is so easy to leave in place for years.

The correction is not difficult and it is not primarily a software problem. It is a coding decision, made once, that client cost advances are posted to an asset account and that reimbursements are posted against that account rather than into fee income. Everything downstream, the ageing list, the write-off, the balance sheet figure, follows from the one entry being right.

Two further details are worth carrying into a conversation with your accountant. The first is that the fee agreement governs the analysis. The IRS guide distinguishes the net fee contract, which “specifically provides that advanced litigation costs are repaid to the attorney before calculating the contingency fee percentage,” from the gross fee contract, under which the percentage is “calculated without regard to advanced litigation costs” and the attorney “is not separately reimbursed.” In the leading appellate case on the point, a firm using a gross fee contract was held to have “incurred ordinary and necessary business expenses in the payment of costs,” the court noting there was “no obligation on the part of the client to repay the money expended.” Our treatment of contingency fee agreements covers the clause that decides this, which is the same clause that decides what the client nets.

The second is that the common argument for deducting anyway has been tested and answered. The guide notes that attorneys often claim costs are currently deductible because there is no guarantee of recovery, and records how that argument fares: “courts specifically look at the attorney’s case selection and fee advancement processes,” so a high rate of recovery “would indicate that the attorney is likely to be reimbursed.” The examiner is instructed to look at your recovery rate. Your own books should tell you what it is before anyone else asks.

The chart of accounts a small firm needs, including the lines no other business has

A chart of accounts is just the list of buckets your transactions can fall into, organised under the five standard classes: assets, liabilities, equity, revenue and expenses. Most small firms accept whatever list their software installed on day one and then spend years categorising into buckets that were designed for a retailer. The Oregon materials name the result as one of the standard pitfalls, “improper chart of accounts setup,” and the consequence as the one you would expect: it “hinders financial analysis and reporting.”

Four of the accounts a law firm needs do not appear on any generic list, and three of them exist only because the firm holds other people’s money.

Two of those rows deserve emphasis because they are the ones that silently corrupt a revenue figure. A firm that routes cost reimbursements through fee income will report revenue it never earned, and will then compare that inflated number to last year, to a benchmark, or to a partner’s origination credit. The Oregon materials list this pitfall in exactly those terms, as a failure of “not separating reimbursable client costs from actual income,” and list its cousin alongside it: “treating loans as income instead of liabilities inflates profits inaccurately.”

The trust pair in the first two rows is the one to set up first, because it is self-checking. If the trust bank account and the client trust liability account do not carry the same figure at the end of the month, the firm has learned something before anyone has had to look for it. That is the whole value of recording the liability: a firm that holds client money and shows only the bank side of it has a balance sheet that quietly claims the money as its own, and no automatic signal when the two sides drift apart.

The same logic applies to the separation of the accounts themselves, which is the physical version of the same idea: two containers, never netted, each with its own records.

Beyond those, the useful design principle is to build the expense side around decisions rather than around receipts. A firm that books everything into “office expense” has records that satisfy a tax preparer and inform nobody. A firm that separates occupancy, insurance including professional liability, technology and subscriptions, staff compensation and the loaded cost of employment, professional development and bar obligations, marketing and client development, and non-reimbursable case costs can look at a statement and see where a decision is available. The number of accounts matters far less than whether each one corresponds to something an owner could choose to change.

Revenue deserves the same treatment. Splitting fee income by practice area, or by fee type where a firm runs more than one, costs nothing at setup and is nearly impossible to reconstruct later. Our guide to law firm profitability covers what to do with that split once you have it.

What the books get wrong most often, and what each error does to the numbers

The Oregon State Bar’s Professional Liability Fund presentation lists eight common pitfalls, and the useful thing about the list is that it is written by people who see the consequences rather than by people selling the cure. Mismanaging trust accounts, through “mixing client funds with firm funds,” writing checks from the wrong account, and failing to address unclaimed funds. Skipping required reconciliations, which “leads to inaccuracies in financial records.” Commingling operating and trust accounts, which the materials attribute specifically to “deposits incorrectly marked into wrong account from payment links, wires.” Failing to separate reimbursable client costs from income. Treating loans as income. Improper chart of accounts setup. Lack of integration between billing and accounting, where “manual data entry causes errors and inefficiencies.” And wire fraud, an “increasing risk with digital transactions without proper safeguards.”

Independent survey evidence points at the same places. A 2025 industry report based on responses from more than 2,800 legal professionals, primarily in solo and small firms, found 68 percent describing fee collection as a significant or moderate challenge, 57 percent reporting that collection of reimbursable expenses is difficult, and 55 percent naming expense tracking as a challenge. The same report put card acceptance at 82 percent of firms, up from 78 percent in 2023, which is worth noticing in this context: payment convenience has improved considerably faster than the recordkeeping behind it.

Where firms say the money handling breaks down

It is worth separating the errors by what they actually damage, because firms tend to treat them all as equally urgent and then fix the easiest one. Some corrupt the profit and loss statement, some corrupt the balance sheet, and a few put the licence at risk whatever the statements say.

The two middle bars are the ones to sit with, because they describe the same failure from two directions. Collecting reimbursable expenses is hard partly because tracking them is hard, and tracking them is hard partly because they were never recorded as anything the firm expects to get back. An advance posted to an expense account generates no reminder, appears on no ageing list and shows up in no receivable total. The firm has turned a debt owed to it into a cost it has already absorbed, and then finds the money difficult to collect, which is not a coincidence.

The card acceptance figure belongs in the same frame. Payments now arrive instantly, from several channels, often without the matter identified in the deposit, and the Oregon materials flag exactly that as a source of commingling: deposits “incorrectly marked into wrong account from payment links, wires.” Speed at the front of the process raises the standard required at the back of it, which is an argument for tightening the coding discipline rather than for slowing the payments down.

Read the last column top to bottom and a pattern appears. Five of these six errors are invisible on the profit and loss statement that most owners actually look at, and visible on the balance sheet that most owners never open. That is the practical argument for reading both every month, and it costs about four minutes.

The close, a calendar rather than a scramble

Books do not go wrong dramatically. They go wrong by a week at a time, until the week is a quarter and the quarter is the reason a return goes on extension. The defence is a cadence written down in advance, assigned to a named person, and short enough that it actually happens.

The Oregon materials set out a workable one and, usefully, name who owns each step. Accounts are set up once, by the attorney and the bookkeeper together. Time and transactions are recorded daily, by the attorney and staff. The three-way reconciliation is monthly, “created by bookkeeper and reviewed by attorney.” Financial reports are reviewed monthly. Financial planning and taxes are discussed quarterly with the tax accountant. And the accounting procedures themselves are reviewed “ongoing and at least annually” by all three.

The cadence only works if each step is small enough to survive a bad week. A close that takes a bookkeeper two days and an owner twenty minutes will happen every month for years. A close that requires reconstructing six weeks of coding will be skipped the first time a trial lands, and then skipped again.

Two items on that list are worth defending against the natural urge to let them slide. The first is the monthly review by the owner rather than by the bookkeeper alone. The reconciliation is a technical exercise and can be delegated. The reading is not, because the bookkeeper cannot know that a particular client’s balance has been flat for four months or that a cost was advanced on a matter the firm has already decided to drop.

There is a related discipline that costs nothing and prevents most of the month-end archaeology. Uncoded transactions should be cleared to zero every week rather than left to accumulate until the close. A transaction coded the week it happened is coded by someone who remembers what it was. A transaction coded five weeks later is coded by someone reading a merchant name on a bank feed, and the result is the undifferentiated “office expense” bucket that makes a statement unreadable. The work is identical in both cases; only the accuracy differs.

Put the cadence somewhere visible rather than in somebody’s memory. A dated checklist that is initialled when each step is done converts the close from an intention into a record, and a record is the thing an owner can check without having to ask.

The second is the quarterly ageing of advanced client costs. An advance recorded correctly as a receivable is still only useful if somebody looks at the list. Costs advanced on matters that resolved two years ago are not receivables in any meaningful sense, and carrying them as assets overstates the firm. Writing them off is a decision, taken once a year, in the open, rather than a discovery made by an accountant in April.

One habit is worth adding to the calendar that the compliance literature does not mention: a quarterly look at the gap between what the firm billed and what it collected. Our treatment of attorney billing covers the cascade from time recorded to cash received, and the point of putting it on the accounting calendar rather than the practice-management one is that the figures already exist in your ledger. Nobody has to build a report. Somebody has to open it.

Reading your own statements, and what the profit line is not

Three statements come out of a properly kept ledger, and each answers a question the other two cannot. The Oregon materials define them economically: the profit and loss shows “how profitable your firm was during a specific time period,” the balance sheet gives an “overview of firm’s assets (own), liabilities (owe), and equity (value),” and the cash flow statement shows “how much cash is coming in and out of your firm during a specific time period.”

Three statements, three questions, three blind spots

Which brings us to the line that causes the most confusion in a small firm, and it is the bottom one. On a firm taxed as a partnership or a sole proprietorship, the profit figure is not owner pay and owner pay is not an expense. The IRS states that a partnership “passes through profits or losses to its partners,” that “each partner reports their share of the partnership’s income or loss on their personal tax return,” and that “partners are not employees and shouldn’t be issued a Form W-2.” A partner’s money reaches them as draws, as guaranteed payments and as a distributive share reported on a Schedule K-1, none of which is a wage.

Two consequences follow for how you read your own statements. The first is that comparing your firm’s profit margin to a published margin for some other kind of business is meaningless unless you know whether owner labour was charged above the line. Two identical firms, one taxed as a partnership and one having elected S corporation treatment, will report materially different profit figures for the same year and the same work, because in the second the working owner’s reasonable compensation is a wage expense and in the first it is not an expense at all. The IRS states the S corporation position directly: “distributions and other payments by an S corporation to a corporate officer must be treated as wages to the extent the amounts are reasonable compensation for services rendered to the corporation.” Our treatment of law firm partner compensation covers those mechanics, and attorney compensation covers what the published wage figures do and do not include.

The second is that the profit line, whatever it says, is not available. Taxes come out of it, capital stays in it, and the remainder is what reaches the owner. A firm that reads profit as take-home will over-distribute in good quarters and discover the shortfall at the first estimated payment. The quarterly step on the calendar above, setting money aside for taxes, exists for exactly this reason, and the third of the three accounts the Oregon materials say every firm needs, a business savings account “for taxes, reserves,” is the place to put it. Opening those three accounts is a day-one decision, covered in our guide to starting a law firm.

What your accountant needs from you, and a self-audit of your own books

The relationship with whoever keeps your books breaks down in a predictable way: the firm sends a bank feed and a shoebox, the bookkeeper makes assumptions, and nobody discovers the assumptions until something does not reconcile. The IRS’s guide for examiners is an unexpectedly good checklist here, because it lists what a well-run law practice has on hand. Attorneys, it says, usually maintain a receipts journal or daily log; a disbursements record that breaks out regular expenses from trust disbursements and allows disbursements chargeable to a specific client to be noted for billing; an accounts receivable journal showing billed receivables; individual client accounts with “a description of services rendered, charges and credits, a summary of unbilled charges and work in progress, and final invoices”; case time records per client; a register of cases in progress; and time summary reports by attorney and by client.

Notice how many of those are about matters rather than about money. A law firm’s books only work when the matter-level records and the general ledger agree, which is why the guide tells examiners to perform a “reconciliation of the output of the time and billing system to the appropriate accounts in the general ledger.” If your billing system and your accounting system have never been reconciled to each other, that is the single highest-value thing on this page.

What follows is a short self-audit. It is not a compliance test and nothing here is legal or tax advice for your firm. It is a list of questions a well-kept set of books can answer in under a minute, and a poorly kept set cannot answer at all.

If most of those are a yes, the books are doing their job and the remaining work is interpretation. If most are a no, the fix is rarely new software. It is usually a chart of accounts rebuilt once, a cadence written down, and one person who is accountable for the close. Our guide to running a law firm treats the financial close as one of the systems that has to be written down rather than carried in somebody’s head, and the software selection question is a separate one, taken after the requirements are known rather than before.

Frequently asked questions