// v4/lesson-3-pages.jsx — Chapter III · Quality of Earnings · all 9 lessons. const V4_CH3_LESSONS = [ ['3.01', 'What QofE actually means', 'lesson-3-01.html'], ['3.02', 'Five categories of normalisation', 'lesson-3-02.html'], ['3.03', 'Owner remuneration', 'lesson-3-03.html'], ['3.04', 'Family on payroll', 'lesson-3-04.html'], ['3.05', 'Personal mix in expenses', 'lesson-3-05.html'], ['3.06', 'Related-party rent', 'lesson-3-06.html'], ['3.07', 'Working capital, normalised', 'lesson-3-07.html'], ['3.08', 'Capex: maintenance vs growth', 'lesson-3-08.html'], ['3.09', 'Off-balance-sheet items', 'lesson-3-09.html'], ]; const ch3 = { chapterRoman: 'III', chapterTitle: 'Quality of Earnings', chapterHref: 'chapter.html', lessonsInChapter: V4_CH3_LESSONS, }; // ============================================================================ // 3.01 — FREE preview · full content // ============================================================================ function V4Lesson301() { const v4 = window.v4; const { V4DropCap, V4H2 } = window; return What "Quality of Earnings" actually means.} brief={<>Reported earnings, sustainable earnings, transferable earnings — three different ideas, all routinely confused. The opening lesson of the chapter that does eighty percent of the work in any serious SME valuation.} readMins={4} body={<> he earnings figure on the financial statements is the number the company reported. The earnings figure a buyer is willing to capitalise is the number the company can be relied upon to produce, after the buyer has bought it. Between the two sits the chapter that almost every serious valuation revolves around. The discipline is called Quality of Earnings — QofE for short — and it accounts for about eighty percent of the work on a typical SME file. Three earnings layers, properly

Most introductions to QofE talk about "normalised earnings." This is too coarse. There are three distinct earnings figures hiding in every SME’s P&L, and confusing them is the source of most disagreements between brokers and buyers.

Reported earnings are what the accountant signed off on. They tie to the bank statement, to the IRD return, to the formal financial statements. They are not in dispute. They are also not particularly useful for valuation, because they reflect the owner’s accumulated decisions about tax and structure, not the underlying earning power of the business.

Sustainable earnings are what the business would earn under continuing-operation assumptions, with all one-off items, non-recurring expenses, and accounting peculiarities removed. They reflect the underlying earning power on a continuing basis, assuming the same operator and the same operating model. This is closer to what a financial buyer might capitalise.

Transferable earnings are what the business would earn under a new owner — accounting not just for one-offs but for the cost of replacing the things the seller did for free, the customer relationships that may not transfer, the supplier terms that depend on personal relationships, and the management overhead that a hired team would impose. This is what a strategic buyer actually capitalises. The gap between sustainable and transferable is often the most contested number in the file.

Why the layers matter for the valuation

A multiple applied to reported earnings produces a number that nobody on either side of the deal will defend. A multiple applied to sustainable earnings produces a number a financial buyer might pay. A multiple applied to transferable earnings produces a number a strategic buyer would. The deal closes at the lower number of these last two, in almost every case.

The valuer’s discipline is to produce all three figures on the page, name them by their proper labels, and show the bridge between each pair. A working file that shows reported EBITDA of $1.2M, sustainable EBITDA of $720k after one-off removal, and transferable EBITDA of $480k after replacement-cost adjustments has done the chapter’s work. A working file that shows "normalised EBITDA of $X" without specifying which of the three layers is meant has skipped the chapter.

The five categories that move the layers

The next lesson (3.02) walks through the five categories of normalisation that take reported earnings to transferable earnings: owner remuneration, family on payroll, personal mix in expenses, related-party transactions, and one-off / non-recurring items. The remaining lessons (3.03 through 3.09) take each in turn, with the seventh category — working capital and capex normalisation — split into its own treatment because it deserves longer.

The order of the chapter is not academic. The categories are walked through in roughly the order they tend to bite in real files: owner remuneration first because it is the biggest single line on most SME P&Ls; family on payroll second because it interacts with owner comp; personal mix third because it is the easiest to defend and the most often disputed; related-party items fourth because they tend to be larger than they look; off-balance-sheet items last because they are the least obvious.

} pullQuote={<>Three earnings layers. Reported is what was reported. Sustainable is what the business earns. Transferable is what survives a sale. The chapter is the bridge between them — and most files close at the third number.} takeaway={<>Reported, sustainable, transferable — three distinct earnings figures, all hiding in the same P&L. The valuer’s discipline is to produce all three on the page and show the bridge between them. The next eight lessons are the bridges.} prevHref={null} nextHref="lesson-3-02.html" nextNum="3.02" nextTitle="The five categories of normalisation" />; } // ============================================================================ // 3.02 — FREE preview · full content // ============================================================================ function V4Lesson302() { const v4 = window.v4; const { V4DropCap, V4H2 } = window; return The five categories of normalisation.} brief={<>Every adjustment that bridges reported to transferable earnings falls into one of five buckets. Naming them properly is the difference between a defended valuation and a list of opinions. The map of the chapter — and the working file the rest of the chapter fills in.} readMins={4} body={<> very normalisation adjustment that a working valuer makes to bridge reported earnings to transferable earnings falls, after enough files, into one of five categories. Naming the category on the page, before writing down the dollar adjustment, is the single biggest defensibility upgrade a valuation report can make. It also produces a working file that survives questioning by a credit committee, a buyer’s lawyer, or a family-court appointed reviewer. 1 · Owner remuneration adjustments

The largest single line on most SME P&Ls. The principal’s salary is rarely a market salary; it is set to balance tax, dividend, and personal liquidity. The replacement-cost framing (lesson 3.03) bridges the reported salary to the market-rate salary for the role the principal is actually performing — which is rarely a single role and often less than the principal believes.

2 · Family on payroll

Spouses, children, parents, and increasingly adult siblings, listed on payroll at non-market rates. Some are doing real work; some are not. The discipline is per-line investigation (lesson 3.04), with a defensible / partially defensible / indefensible classification for each name, and a market-rate substitution where the line stays in.

3 · Personal expenses through the company

Three sub-categories that show up on almost every SME P&L: the personal vehicle, the personal travel, and the personal credit-card mix (lesson 3.05). Each is straightforward to bound; each is routinely disputed by sellers in negotiation; each is unambiguous in the valuer’s working file.

4 · Related-party transactions

The largest single hit to sustainable earnings on most owner-occupied SME files. Below-market rent on owner-occupied premises (lesson 3.06) is the canonical case; below-market loans, royalty-stripping arrangements, and family-member contracts at non-arm’s-length terms follow the same logic. The arm’s-length test is not a checkbox; it is a defended comparison.

5 · One-off and non-recurring items

The category most sellers want to use aggressively and most buyers want to disallow. The COVID-period bounce, the unusual product-line write-off, the litigation settlement, the one-time IT investment. The discipline is symmetry (lesson 3.09): a true one-off should be added back if expense, deducted if income — both directions, applied equally.

The working file structure

A well-built QofE working file is structured around these five categories. The file has a tab for each, and within each tab a numbered line for every adjustment, with a "defensible / partially defensible / indefensible" column, a dollar amount column, and a "comments" column where the defence is written out. A reviewer looking at the file should be able to follow the logic line by line without asking questions.

Two additional treatments — working capital and capex — are technically not normalisations of reported earnings; they are sustainability checks on the cash-flow conversion. They get their own lessons (3.07 and 3.08) and a separate tab in the working file. Off-balance-sheet items (3.09) are technically not earnings either, but they belong in the discussion because they routinely change the deal economics.

} pullQuote={<>Five categories, in order of how often they bite: owner comp, family payroll, personal mix, related-party, one-offs. Naming the category before writing the dollar adjustment is the biggest single defensibility upgrade a valuation report can make.} takeaway={<>The five buckets: owner remuneration · family payroll · personal mix · related-party · one-off / non-recurring. Each gets its own lesson in this chapter. Plus working capital and capex (3.07-3.08) and off-balance-sheet (3.09) as adjacent disciplines.} prevHref="lesson-3-01.html" prevNum="3.01" prevTitle="What QofE actually means" nextHref="lesson-3-03.html" nextNum="3.03" nextTitle="Owner remuneration: replacement-cost framing" />; } // ============================================================================ // helper for locked lessons // ============================================================================ function makeLocked3({ num, title, brief, readMins, intro, memberBody, pullQuote, takeaway, lockedBundled, prevHref, prevNum, prevTitle, nextHref, nextNum, nextTitle }) { return function Lesson() { return ; }; } const HL = { background: '#f2e600', padding: '0 10px', boxShadow: `4px 4px 0 #0d0d0d` }; // ============================================================================ // 3.03 — Owner remuneration // ============================================================================ const V4Lesson303 = makeLocked3({ num: '3.03', title: <>Owner remuneration., brief: <>Replacement-cost framing. The principal’s reported salary is rarely a market salary, and the role they perform is rarely a single role. The largest single line on most SME P&Ls — and the one most commonly bungled by analysts who skip directly to a "market replacement salary" without thinking about what is actually being replaced., readMins: 5, intro: <> {(() => { const { V4DropCap, V4H2 } = window; const v4 = window.v4; return <> he principal of a closely-held SME pays themselves an amount that has almost nothing to do with the market rate for the role they perform. It is set to balance tax minimisation against personal liquidity against the imputation regime against the demands of a mortgage covenant. It is, in short, a personal-finance number. For a valuer it is a placeholder, to be replaced — but replaced with what? The replacement-cost framing

The conceptual frame is to ask: what would the business have to pay, at open-market rates, to replace the functions the principal is performing? This is the right starting point, but it is rarely the right ending point. The principal of a small distribution business may be variously the sales director, the credit controller, the head of buying, and the head of compliance. Replacing all four functions one-for-one is rarely what a buyer will actually do.

; })()} , memberBody: <> {(() => { const { V4H2 } = window; const v4 = window.v4; return <> What a buyer actually intends to do

The replacement-cost framing only works if you can specify, in advance, the buyer’s intention for the role. A strategic buyer who is folding the business into existing infrastructure typically replaces with one mid-level operator and absorbs the rest into shared services. A financial buyer who is keeping the business operating standalone typically replaces with a full hired team. An owner-operator buyer typically replaces one-for-one.

These three buyer types produce different replacement costs and therefore different transferable EBITDA figures. The working valuer’s discipline is to do the replacement at each of the three buyer-type intentions and present the resulting range. Brokers routinely pick the lowest replacement cost (which produces the highest transferable EBITDA) and present it as "the" number. Buyers routinely pick the highest. The valuer’s file shows all three.

The four-function decomposition

Before pricing a replacement, decompose the principal’s actual work into the four functions they perform. Use the principal’s own calendar from the last quarter; this is more honest than asking them. Categorise each block of time as sales, operations, finance/admin, or strategic / external. Then ask, for each function, whether the buyer’s intended team has equivalent capacity already.

We routinely find that 35-50% of a principal’s hours are in functions that a strategic buyer will not need to replace at all (the relationship-driven sales work transferring to an existing salesperson; the operations work being absorbed by a regional manager). Pricing the replacement at full market rate for all four functions overstates the cost — and produces a transferable EBITDA below what the strategic buyer actually faces.

The transition allowance

Even where a function will not be replaced long-term, there is usually a transition period during which both the principal and the buyer’s replacement are needed. Twelve months is the modal answer; six is aggressive; twenty-four is occasionally defensible for relationship-heavy businesses. The transition cost is a real adjustment to year-one cash flow but should not be capitalised — it is a one-time cost, not a continuing expense.

The mistake to avoid is double-counting. If the transition allowance includes paying the principal an earn-out, and the earn-out is also being treated as part of the deal consideration, you have priced the same dollar twice. The working file separates "transition cost in normalised EBITDA" (zero — it is one-off) from "transition consideration in deal structure" (positive — it appears in the purchase price waterfall).

The Case 047 worked example

On the Auckland import-trading SME that opened the chapter, the two principals had been paying themselves $80,000 each. The replacement-cost framing produced three candidates: $440,000 (full replacement at market for all four functions, two principals); $260,000 (one combined mid-level operator under a strategic buyer assumption); $180,000 (single operator with significant function absorption). The buyer’s analyst — correctly, in our reading — used the middle figure. The adjustment to transferable EBITDA was a $100k reduction from the reported $80k + $80k = $160k owner comp.

; })()} , pullQuote: <>Owner remuneration is not "owner’s actual pay" minus "market salary." It is the cost of the functions the principal performs, replaced at the level the buyer actually intends to replace them. Specifying the buyer’s intention is the work., takeaway: <>Decompose the principal’s work into four functions. Price each at the buyer’s intended replacement level (strategic / financial / owner-operator). Show all three candidates in the file. Treat transition cost as one-off, not as continuing., prevHref: 'lesson-3-02.html', prevNum: '3.02', prevTitle: 'The five categories of normalisation', nextHref: 'lesson-3-04.html', nextNum: '3.04', nextTitle: 'Family on payroll', }); // ============================================================================ // 3.04 — Family on payroll // ============================================================================ const V4Lesson304 = makeLocked3({ num: '3.04', title: <>Family on payroll., brief: <>Spouse, child, parent: defensible, partially defensible, indefensible. The per-line investigation that produces a column titled "defensible" in your working file — and the four signals that classify a name into one of the three buckets without a long conversation., readMins: 4, intro: <> {(() => { const { V4DropCap, V4H2 } = window; const v4 = window.v4; return <> s often as not, an owner-run SME P&L carries at least one family member on the payroll beyond the principal. Of those, perhaps a third are doing real work at defensible rates; another third are doing real work at non-market rates; the final third are not doing meaningful work at all. The valuer’s discipline is per-line investigation: every family name on the payroll gets its own treatment. The three buckets

Each family name lands in one of three buckets. Defensible: doing real work, at or near market rate. The line stays as-is. Partially defensible: doing real work, but at a non-market rate (either above or below). The line stays in, adjusted to market. Indefensible: not doing meaningful work, or doing work that a buyer would not retain. The line comes out entirely.

; })()} , memberBody: <> {(() => { const { V4H2 } = window; const v4 = window.v4; return <> The four signals

Four diagnostic signals classify a name quickly. (1) Hours actually worked. The principal’s own estimate is unreliable here; ask for the family member’s calendar, email volume, and (where possible) the team’s view of what they do. (2) Reporting structure. Does someone manage them? Do they manage someone? An organisational dotted-line to no one is a strong negative signal. (3) Function uniqueness. Is the work they do work that someone else in the business cannot do, or duplication? (4) Market comparability. If this role were hired externally, what would it pay — and would it be hired at all?

Two or more positive signals: defensible. One: partially defensible, normalise to market. Zero: indefensible, remove.

The spouse case

The spouse on payroll, listed as "administrator" or "operations manager," is the canonical case and the one most often disputed in negotiation. About half the spouses we examine are doing real administrative work at non-market rates (typically over-paid relative to the function, occasionally under). The other half are nominal — the line exists for tax structure, not for work performed.

The defence in negotiation is straightforward when the working file is honest: a spouse at $80,000 doing eight hours per week of bookkeeping work is being paid roughly four times market for the role; the adjustment is $60,000 down. A spouse at $80,000 doing four days per week of customer-facing sales work is probably under-paid; the adjustment might be $20,000 up.

The children case

Adult children on payroll at "sales associate" or "associate manager" titles. The full spectrum turns up: highly capable adults doing real sales work at appropriate rates; somewhat-capable adults doing some work at premium rates; nominal listings where the child has not been on premises for two years. The four-signal test handles all three honestly.

The disputed case is the in-between: the adult child doing some work, at a premium rate, with limited management oversight. A sound working assumption is to normalise to market for the work actually performed — which often produces a 40-60% reduction in the payroll line.

The cross-border family case

Where the family member is non-resident, the analysis tightens. A Hong Kong-resident director who has not visited New Zealand in three years, drawing $120,000 per annum as a director’s fee, is not being paid for work — they are being paid through a structure. Treat the line as a remittance, not an expense, and remove from the operating P&L entirely. This becomes a Chapter VI cross-border issue (lesson 6.07) and is often the single largest cross-border QofE adjustment.

; })()} , pullQuote: <>Defensible, partially defensible, indefensible. Four signals, three buckets, one column in the working file. The spouse listed as administrator is the canonical case — about half are real, about half are not. Both halves get the same diagnostic., takeaway: <>Per-line investigation for every family name. Four signals (hours, reporting structure, function uniqueness, market comparability) classify into three buckets. Defensible stays; partially defensible normalises; indefensible comes out. Cross-border family members are usually remittance, not expense., prevHref: 'lesson-3-03.html', prevNum: '3.03', prevTitle: 'Owner remuneration', nextHref: 'lesson-3-05.html', nextNum: '3.05', nextTitle: 'Personal mix in expenses', }); // ============================================================================ // 3.05 — Personal mix // ============================================================================ const V4Lesson305 = makeLocked3({ num: '3.05', title: <>The car, the trip, the credit card., brief: <>Three sub-categories of personal expense that show up on almost every SME P&L. Each is straightforward to bound; each is routinely disputed in negotiation; each is unambiguous in the working file once the rule is set., readMins: 4, intro: <> {(() => { const { V4DropCap, V4H2 } = window; const v4 = window.v4; return <> f the five categories of normalisation, this one is the most-disputed in negotiation and the easiest to bound in the working file. Personal expenses run through the company P&L fall into three reliable sub-categories: vehicles, travel, and credit-card mix. Each has its own rule. The principal is rarely happy about any of them in negotiation. None of them, in the working file, is ambiguous. 1 · The car (or two)

The leased or owned vehicle for predominantly personal use is the canonical case. The accounting trail is clean — there is a lease or a depreciation schedule on the books — and the use case is usually mixed. The rule: if the vehicle is plausibly required for the business (a delivery van, a sales rep’s car), it stays; if it is plausibly personal (a Range Rover Sport, a luxury sedan), it comes out at fair use rate.

; })()} , memberBody: <> {(() => { const { V4H2 } = window; const v4 = window.v4; return <>

On owner-occupied SMEs, the principal typically has two vehicles run through the company — one ostensibly for business, one for personal. Both are usually expensed. The adjustment is, at minimum, to remove one entirely (the second is unambiguously personal). The first is often partial; a reasonable default is a 50% personal-use add-back on the working vehicle’s costs in the absence of a defensible logbook.

2 · The trip

Personal travel booked through the company. The classic forms: a family holiday described as a "trade show attendance," a personal trip combined with a brief client visit, an annual flight to Europe described as "supplier negotiations." The defence the seller mounts in negotiation is usually circumstantial — there was a meeting, there was a coffee, there was a contact made.

The rule: was the trip booked through the company in a form a buyer’s management team would book it? A three-week family holiday with one client meeting is not. A two-day flight to Auckland from Wellington for a half-day of client visits is. The grey middle (a one-week trip with two days of meetings) is split pro-rata between business and personal.

3 · The credit card mix

The widest category and the most-disputed. Personal grocery purchases, family dinners, mobile phones, broadband, school fees, family medical, gym memberships, club fees — all routinely run through the company credit card. The pattern is large per-month, small per-line, hard to defend in detail.

The rule: pull a sample of three months of credit-card statements, categorise each line manually, and produce a personal-use ratio. The ratio is then applied to the full year’s credit-card spend. Do not try to argue every line individually — that is the broker’s strategy and it loses. We argue the ratio.

The size of the line, in practice

For an owner-occupied SME with reported EBITDA of $1.2M, the personal-mix adjustment routinely sits between $40,000 and $120,000. On Case 047 it was $95,000 — the Range Rover lease at $34k, personal travel at $18.6k, and family medical at $42.1k.

The category is also where the working file’s "comments" column earns its keep. Each line has a defended classification. A reviewer with the working file can audit every dollar; a reviewer with just a number must take the valuer’s word for it. The first is professional work; the second is not.

; })()} , pullQuote: <>Three sub-categories: vehicles, travel, credit-card mix. Each has its own rule. Each routinely contributes $30-50k of adjustment on an owner-occupied SME. None is ambiguous in the working file — only in the negotiation that follows., takeaway: <>Vehicles: defensible business use stays, personal comes out. Travel: book-form test, with pro-rata for mixed trips. Credit-card mix: three-month sample, manual categorisation, applied ratio. Total adjustment typically $40-120k for an owner-occupied SME., prevHref: 'lesson-3-04.html', prevNum: '3.04', prevTitle: 'Family on payroll', nextHref: 'lesson-3-06.html', nextNum: '3.06', nextTitle: 'Related-party rent', }); // ============================================================================ // 3.06 — Related-party rent // ============================================================================ const V4Lesson306 = makeLocked3({ num: '3.06', title: <>Related-party rent., brief: <>The single largest adjustment in most owner-occupied SME files. Arm’s-length is a test, not a checkbox. Building a defensible comparable for property leases — and what to do when the rent is genuinely market but the lease terms favour one side., readMins: 4, intro: <> {(() => { const { V4DropCap, V4H2 } = window; const v4 = window.v4; return <> he single largest hit to sustainable EBITDA on most owner-occupied SME files. The pattern: a trading company that pays rent to a family-owned property holding company for the warehouse, factory, or office premises. The rent is rarely arm’s-length. Below-market rent flatters trading EBITDA; above-market rent does the opposite. Brokers and accountants treat this line inconsistently; the working valuer must treat it systematically. The arm’s-length test

Arm’s-length is not a single number. It is a range — what would the trading company be paying if it were leasing from an unrelated third-party landlord, on market terms, in the same property, in the same condition, with the same lease structure? The range is real because real third-party leases come in a spread; pretending to a point estimate is dishonest.

; })()} , memberBody: <> {(() => { const { V4H2 } = window; const v4 = window.v4; return <> Building the comparable

Three sources build a market-rent comparable: the public real-estate database (Realestate.co.nz commercial), the most recent two-to-three years of lease registrations from the LINZ database, and — where the building is unusual — an in-person comparable visit to three to five similar-sized facilities in the same area. The output is a range, typically with a 25-30% spread, from which the median is used as the working figure.

The comparable must match the property in three ways: location (same industrial estate or commercial precinct), size (within ±25% of the trading company’s footprint), and condition (the trading company has often invested in fit-out the property holdco did not pay for; this counts as part of the rent if the fit-out stays with the building on lease termination).

The fit-out trick

Many owner-occupied SMEs have invested heavily in fit-out of their premises — racking, cool-rooms, office partitioning — which has been expensed by the trading company but never compensated by the property holdco. On a new third-party lease, the same fit-out would either have been paid by the landlord (and recovered through a higher rent) or by the tenant (and amortised over the lease term).

The working valuer’s discipline is to add the imputed fit-out cost to the rent comparison. A trading company paying $90k rent on a building where the third-party rent would be $250k, with $60k of unamortised fit-out per annum, has a true arm’s-length rent of $310k. The adjustment is $220k, not the $160k a casual reader might infer.

The lease-term issue

Even where rent is arm’s-length, the lease terms often favour the related party. Long-term leases with no rent reviews advantage the tenant in a rising market; short-term leases with the option to terminate advantage the landlord. The valuer’s working file should note the lease term separately from the rent figure — the buyer will, on completion, almost certainly be required to enter a new lease at market.

The Case 047 number

On the Auckland import-trading SME: warehouse rent was $90k against a market rent (median of three comparables, plus $45k of unamortised fit-out) of $315k. The adjustment was $225k — the single largest line in the QofE working file, and (in the brokers’ eventual concession) the line that closed the gap from "incomprehensible offer" to "defensible offer."

The seller’s argument was that the related-party rent had been in place for ten years and was therefore "market" by usage. This is incorrect. Market is what a new lease, today, between independent parties, would price at. A long history of below-market lease arrangements does not establish market; it establishes a long history.

; })()} , pullQuote: <>The largest single adjustment on most owner-occupied SME files. A long history of below-market rent does not establish market — it establishes a long history. Market is what a new lease, today, between independent parties, would price at., takeaway: <>Build the comparable from three sources, match on location/size/condition, use the median. Add imputed fit-out cost to the rent comparison. Note lease terms separately. The adjustment is routinely 100-300% of the reported rent line and frequently dominates the QofE file., prevHref: 'lesson-3-05.html', prevNum: '3.05', prevTitle: 'Personal mix in expenses', nextHref: 'lesson-3-07.html', nextNum: '3.07', nextTitle: 'Working capital, properly normalised', }); // ============================================================================ // 3.07 — Working capital // ============================================================================ const V4Lesson307 = makeLocked3({ num: '3.07', title: <>Working capital, normalised., brief: <>Not just average days. The mid-cycle problem, the seasonality problem, the supply-shock problem. Working capital is not an earnings adjustment in itself, but its movements distort the cash-flow conversion and routinely cost SME buyers six-figure surprises in year one., readMins: 4, intro: <> {(() => { const { V4DropCap, V4H2 } = window; const v4 = window.v4; return <> orking capital normalisation is the QofE chapter’s longest lesson because it is the discipline most analysts skim and the source of the largest surprises in year one of a new ownership. Working capital is not an earnings adjustment — it does not change EBITDA — but its movements change the cash the buyer actually receives, and getting the normalisation wrong is the difference between a buyer who closes happy and a buyer who comes back six months later asking for an earn-out adjustment. The three problems

Three structural problems with the conventional "average days outstanding" approach to working capital normalisation. The first is the mid-cycle problem: a year-end snapshot taken at a non-representative moment in the business’s cycle. The second is the seasonality problem: a business whose working capital varies by 30-40% across the year, where annual averages mislead. The third is the supply-shock problem: a business that has temporarily extended payables or run down inventory in response to specific recent events.

; })()} , memberBody: <> {(() => { const { V4H2 } = window; const v4 = window.v4; return <> The mid-cycle problem

Most SME accounts close on 31 March (NZ tax year) or 30 June. Both dates are arbitrary from a working-capital perspective. A retail business with a Christmas peak will have very different working capital at 31 March (post-peak, inventory low, receivables high) versus 30 September (pre-peak, inventory building, receivables low). A snapshot at either date misrepresents the mid-cycle position.

The correction is to take twelve monthly snapshots through the year, average them, and use the average as the normalised working capital level. This requires monthly trial balances, which most SMEs do not produce; we typically need to reconstruct from GL extracts. The exercise takes a working day for a smaller SME and is the largest single time investment in a QofE engagement.

The seasonality problem

Heavily seasonal businesses (retail, agriculture, hospitality, parts of construction) have working capital swings of 30-50% across the year. Using annual averages here is dangerous because the year-end and the operational mid-point can be far apart, and the buyer is going to be financing the working capital expansion that happens mid-year.

The discipline is to model two working-capital figures: the mid-cycle level (used for the long-run sustainable conversion) and the peak-cycle level (used for the year-one financing requirement). The gap between the two is what the buyer needs from a working-capital facility on day one. Sellers routinely understate this; buyers routinely discover it three months into ownership and become unhappy.

The supply-shock problem

Recent events — COVID supply disruption, container-rate volatility, the 2022-23 inventory glut — have caused many SMEs to temporarily change their working-capital posture. A business that ran down its inventory by 35% during 2023 to manage cash, and is now in the process of rebuilding it, has a working-capital position that is not its long-run sustainable position.

The valuer’s discipline is to identify the pre-shock baseline (typically the 2018-2019 working-capital level, adjusted for revenue scale) and use that as the normalised figure. The current trailing-twelve-months figure is wrong in either direction, depending on where in the recovery the business sits.

The earn-out implication

Most SME sale agreements now carry a working-capital adjustment in the SPA — a delivered-working-capital target with a true-up mechanism. Getting the target wrong by 10% of trading capital can result in a six-figure post-completion adjustment. On a $4M SME deal, the working-capital true-up can reach $400k in the buyer’s favour, six months after close.

The discipline here is not just to normalise — it is to specify the normalised figure in language the SPA can adopt. The QofE working file produces both the normalised number and the methodology by which it was derived, in a form the buyer’s lawyers can write into the agreement.

; })()} , pullQuote: <>Working capital is not an earnings adjustment — but its movements change the cash the buyer actually receives. The three problems — mid-cycle, seasonality, supply-shock — turn a tidy P&L into a six-figure surprise three months after close., takeaway: <>Twelve monthly snapshots average to mid-cycle. Seasonal businesses need both mid-cycle (for sustainable) and peak (for financing). Supply-shocked businesses use pre-shock baselines, adjusted. The discipline produces both a normalised number and SPA-ready language., prevHref: 'lesson-3-06.html', prevNum: '3.06', prevTitle: 'Related-party rent', nextHref: 'lesson-3-08.html', nextNum: '3.08', nextTitle: 'Capex: maintenance vs growth', }); // ============================================================================ // 3.08 — Capex // ============================================================================ const V4Lesson308 = makeLocked3({ num: '3.08', title: <>Capex: maintenance vs growth., brief: <>The split most analysts get wrong, and why it changes both free cash flow and the multiple you can defend. Three ways to identify maintenance capex when management cannot — and the bracket-creep mistake that systematically over-states it., readMins: 4, intro: <> {(() => { const { V4DropCap, V4H2 } = window; const v4 = window.v4; return <> f all the SME normalisations, the capex split is the one most analysts get wrong, and the one with the largest second-order effect. The split between maintenance capex (the spending required to keep the business running at its current level) and growth capex (the spending that funds expansion) determines both free cash flow and, indirectly, the multiple a buyer can defend. Mis-split and you produce a valuation that does not survive the buyer’s diligence. Why management can’t tell you

Asking management to split historical capex between maintenance and growth is almost never useful. They have not been keeping records in that form; their split, when they produce one, is reverse-engineered to justify the company’s preferred narrative. The working valuer has to do the split independently from the underlying capex schedule.

; })()} , memberBody: <> {(() => { const { V4H2 } = window; const v4 = window.v4; return <> The three identification methods

Method one: depreciation-anchored. Take the trailing five-year average depreciation as a baseline for maintenance capex. This works for businesses with stable asset bases — a small manufacturer, a logistics business with a fleet. It does not work for technology-heavy businesses where the depreciation schedule reflects accounting choices more than economic life.

Method two: asset-class walkdown. Decompose the asset base into classes (vehicles, plant, IT, fit-out, intangibles), assign an economic life to each, and compute the implied maintenance capex as the annual replacement cost. This is more work but produces a more defensible answer for businesses where method one fails.

Method three: capex-history regression. Plot historical capex against revenue and identify the intercept and slope. The intercept (capex at zero revenue) approximates maintenance for the existing asset base; the slope (capex per dollar of revenue) approximates the growth-related capex. This works well for businesses with at least five years of clean capex history.

The bracket-creep mistake

The most common error in capex normalisation is bracket creep — categorising as "growth" any capex that produced future-period benefit. By this logic, almost every capex item is growth capex, since the new vehicle is for the next five years, the new piece of equipment is for the next eight years, and so on. This systematically over-states growth capex and produces a maintenance figure too low to be credible.

The corrective is to ask the right question: was the business capable of producing the same revenue without this capex? If yes (the new vehicle replaced an old vehicle; the new piece of equipment replaced a broken one), it is maintenance. If no (the new vehicle was the fourth in a three-vehicle fleet; the new equipment opened a new product line), it is growth. This test is binary and produces defensible answers in 80-90% of capex lines.

The case-by-case numbers

On Case 047, reported total capex over the trailing three years averaged $185k per annum. Management proposed a $135k maintenance / $50k growth split. The asset-class walkdown produced $95k maintenance / $90k growth; the depreciation anchor produced $110k maintenance; the regression produced $100k maintenance / $85k growth. We used $100k as the working figure — defensibly between the three methods, materially lower than management’s proposal.

The $35k gap between management’s number and ours moves free cash flow by $35k per annum. At a 4.5× multiple — even if we used the broker’s number — that is $157k of valuation. The capex split, often treated as a footnote, is in fact one of the largest single drivers of the valuation conclusion.

; })()} , pullQuote: <>Management’s split is reverse-engineered to support their narrative. Three independent methods — depreciation-anchored, asset-class walkdown, capex-history regression — produce a defensible answer the buyer can audit., takeaway: <>Don’t accept management’s split. Use the three methods (depreciation, asset-class, regression). The right question: would the business have produced the same revenue without this capex? Binary test, defensible answer, large effect on valuation., prevHref: 'lesson-3-07.html', prevNum: '3.07', prevTitle: 'Working capital, normalised', nextHref: 'lesson-3-09.html', nextNum: '3.09', nextTitle: 'Off-balance-sheet items', }); // ============================================================================ // 3.09 — Off-balance-sheet // ============================================================================ const V4Lesson309 = makeLocked3({ num: '3.09', title: <>Off-balance-sheet items., brief: <>The hidden liabilities. IP held personally. Personal guarantees. Operating leases that pre-date IFRS 16. Contingent payments. The closing lesson of the chapter — the items that do not show up on the balance sheet but routinely change the deal economics., readMins: 4, intro: <> {(() => { const { V4DropCap, V4H2 } = window; const v4 = window.v4; return <> he chapter closes with the items that do not show up on the balance sheet but routinely change the deal economics. These are not earnings adjustments — they are claims against the assets and liabilities. The QofE working file has them in their own section because they belong in the discussion: the buyer needs to know what they are signing up for, and the price the buyer is willing to pay reflects what is taken on as well as what is bought. Four categories

Four categories of off-balance-sheet item show up regularly. Personal IP: trademarks, domains, patents held by the principal personally rather than the company. Personal guarantees: bank facilities, lease guarantees, supplier-credit guarantees personally provided by the principal. Operating leases: lease commitments that pre-date IFRS 16 adoption and have not been capitalised. Contingent payments: earn-outs to previous owners, deferred consideration, contingent royalties.

; })()} , memberBody: <> {(() => { const { V4H2 } = window; const v4 = window.v4; return <> Personal IP

Trademarks and domains held personally by the principal are remarkably common — typically because they were registered before the company was incorporated, or because the principal’s accountant advised on holding them outside the trading entity for tax reasons. The discipline is to confirm transfer on close, in the SPA, at no incremental cost. Deals do get renegotiated mid-process when the trademark turns out, on inspection, not to be included.

Personal guarantees

The bank facilities, lease arrangements, and supplier credit lines that have been personally guaranteed by the principal will need to be re-guaranteed by the buyer on close. The buyer’s discount for taking on the guarantee is real — typically 0.5-1.5% of the underlying facility size, capitalised over the remaining term. For a $1.5M working-capital facility with a five-year term, the implicit cost to the buyer is $40-110k.

Operating leases pre-IFRS 16

Smaller NZ SMEs adopted IFRS 16 inconsistently. Many continue to expense lease payments as operating expenses without capitalising the lease commitment. This understates both balance-sheet liabilities and (because depreciation replaces lease expense) EBITDA in the buyer’s post-acquisition view. The discipline is to compute the IFRS 16-adjusted EBITDA alongside the reported number for any business with material lease commitments.

Contingent payments

Earn-outs to previous owners, contingent consideration on prior acquisitions, royalty arrangements with terminating dates — all routinely missed by buyers in diligence and routinely undisclosed by sellers in disclosure schedules. The working valuer’s discipline is to read the past three years of management accounts looking for "other operating expense" or "professional fees" lines large enough to mask a contingent payment, and to ask directly about prior acquisitions and any continuing obligations.

The chapter’s closing argument

Quality of Earnings is not, in the end, about earnings. It is about producing a defensible bridge between what the financial statements say and what the buyer is actually buying. The five normalisation categories (lessons 3.02 through 3.06) do most of the visible work. Working capital, capex, and off-balance-sheet items (3.07 through 3.09) are where the post-completion surprises live. A QofE engagement that covers only the five categories and skips the last three has done the visible 60% of the work and left the buyer to discover the other 40% on their own time.

The chapter closes here. Chapter IV is the multiples chapter — the multiplier applied to the QofE-normalised earnings to produce the valuation conclusion. The two chapters together do, on most files, ninety percent of the methodology work the valuer is asked to do.

; })()} , pullQuote: <>QofE is not, in the end, about earnings. It is about producing a defensible bridge between what the statements say and what the buyer is actually buying. The last three lessons are where the post-completion surprises live., takeaway: <>Four categories: personal IP, personal guarantees, pre-IFRS 16 operating leases, contingent payments. None show up on the balance sheet; all change the deal economics. The QofE working file’s closing section. Then on to Chapter IV — the multiples that apply to all of this., prevHref: 'lesson-3-08.html', prevNum: '3.08', prevTitle: 'Capex: maintenance vs growth', nextHref: null, }); Object.assign(window, { V4Lesson301, V4Lesson302, V4Lesson303, V4Lesson304, V4Lesson305, V4Lesson306, V4Lesson307, V4Lesson308, V4Lesson309, });