Net zero accounting: what a credible target requires from your ledger

Net zero accounting is the measurement discipline behind a net zero claim: a frozen baseline, a defined boundary, a tracked reduction pathway, and a small, explicit residual balanced by durable carbon removals. A net zero target is an accounting commitment before it is anything else, which most companies discover eighteen months after the press release, when someone asks which baseline year the promised 50% cut is measured against and the honest answer is "it depends which spreadsheet".

This post walks through the five accounting decisions inside a credible target: baseline, boundary, pathway, residual, and claims language. None of them requires an oath to any particular framework, though the Science Based Targets initiative's Corporate Net-Zero Standard shapes the mainstream expectations, and this post reflects those expectations without pretending they are law.

Decision 1: the baseline year, frozen and recomputable

Every reduction claim is relative to a baseline, so the baseline is the most load-bearing number in the whole program. Freezing it means more than writing down a total: it means keeping the ledger, methods and factors that produced it, so it can be recomputed. Because it will need recomputation: baselines must be restated when structure changes materially (acquisitions, divestments), when methods improve (spend-based lines upgraded to activity data), and when errors surface. A baseline that cannot be recomputed cannot be restated, and then every subsequent-year comparison quietly becomes fiction. This is climate accounting in the most literal sense: the practice is called restatement in financial reporting, and it needs the same machinery here, versioned methods, explicit restatement events, an audit trail. That machinery is the core of our carbon accounting software, and it is the difference between "our 2024 baseline, restated for the 2026 acquisition, per the change log" and an argument nobody can win.

Decision 2: the boundary, all three scopes, no creative exclusions

Mainstream net zero expectations require the target to cover the full value chain: Scopes 1, 2 and 3, with Scope 3 typically demanded when it exceeds 40% of the total, which it does for nearly everyone (see why Scope 3 dominates). This is where target design meets inventory quality: you cannot commit to cutting category 1 or category 11 by 90% if your inventory estimates them from a single spend factor with no supplier detail. The dependency runs one way: inventory maturity first, target credibility second. A company that announces net zero before it can produce a defensible category-level Scope 3 number has committed to managing what it has not measured.

Decision 3: the pathway, deep cuts, not purchased silence

The arithmetic of a mainstream net zero pathway is blunt: roughly 90% absolute reduction against the baseline by the target year, with offsets playing no role in the reduction claim itself. Reductions must show up in the inventory as fewer gross tonnes, category by category, year over year. Accounting-wise this creates the reporting obligation companies underestimate most: the annual inventory must stay method-consistent enough that a 4% year-over-year reduction is signal rather than factor-set noise. When a factor library updates, and they update constantly, the change must be isolated and disclosed, or your trend line is measuring the factor vendor's publishing schedule.

Decision 4: the residual and removals, small, explicit, never netted quietly

Net zero does not mean zero: it means residual emissions, on the order of the last 10%, balanced by durable carbon removals. The accounting rules that keep this honest are simple to state. Gross emissions and removals are reported separately, never silently netted. Removals must be removals (carbon physically drawn down and stored), not avoided-emission credits, and their durability should be stated. And until the target year, buying credits is optional decoration on top of reductions, not a substitute for them. The inventory side of this is one line of discipline: offsets and removals live outside the gross inventory, in their own register, with their own evidence.

Decision 5: the claims language, matched to the ledger

Regulators now police green claims (the EU's directives on empty environmental claims being the sharpest example), and the safe harbour is boring precision. "Net zero by 2040 across Scopes 1, 2 and 3, 90% absolute reduction from a 2024 baseline, residual balanced by durable removals, progress reported annually" is a claim an auditor can test against a ledger. "Carbon neutral since 2022" backed by cheap avoidance credits is the kind of claim that now generates enforcement letters. The rule of thumb: never publish a climate claim your emissions ledger cannot mechanically substantiate, because the gap between marketing and ledger is exactly where the legal risk lives.

Net zero, carbon neutral and climate positive are not synonyms

The three phrases circulate as if interchangeable, and they describe different accounting constructions. Carbon neutral, as commonly used, means that the emissions inside some chosen boundary were balanced by purchased credits in a single year. The boundary is frequently Scopes 1 and 2 only, the credits are usually avoidance credits rather than removals, and no reduction is required: a company can grow its gross emissions every year and stay carbon neutral indefinitely by buying more certificates. Net zero, in the mainstream corporate sense, is a long-horizon commitment across the full scope 1 2 3 emissions boundary: deep absolute cuts first, with only a small residual balanced by durable removals at the target year. Climate positive, sometimes styled carbon negative, claims to remove more than the residual, and it is the least standardized of the three, which makes it the most dangerous phrase to print.

The distinction has stopped being rhetorical. Consumer protection regulators in the EU and elsewhere now treat generic environmental claims, and especially offset-backed neutrality claims, as actionable when the substantiation is thin. For an accountant the practical translation is that each phrase implies a different evidence package. A carbon neutral claim needs a complete gross inventory for the year plus a credit register reconciled against it. A net zero target needs a frozen baseline, a reduction pathway and a decade or more of method-consistent annual closes. A climate positive claim needs all of that plus verified removals exceeding the residual, audited to a standard that barely exists yet. If the ledger cannot mechanically produce the evidence package for a phrase, the company should not publish that phrase. Choosing claims language is a disclosure decision, not a branding decision, and it belongs with the people who sign disclosures.

The interim target problem

A 2040 or 2050 net zero year is comfortably abstract. The near-term target is not, and mainstream expectations require one: an interim absolute reduction around 2030, commonly in the region of 42% for a 1.5 degree aligned pathway on Scopes 1 and 2, with Scope 3 targets set alongside. Spread over the years between a 2024 baseline and 2030, that is roughly 4 to 5% of absolute reduction every single year, through acquisitions, energy price swings, factor library updates and whatever the business decides to do commercially. The interim target is where net zero stops being a communications artefact and becomes a recurring reporting obligation with a short deadline.

The accounting consequence is annual close discipline. Six reporting cycles separate a 2024 baseline from a 2030 target, and each one has to be produced on the same boundary, with restatements applied consistently and factor updates isolated from real change, or the trend line is unusable. Miss the discipline for two years and the arithmetic turns hostile: if year three shows 9% cumulative reduction where the pathway wanted 18%, the remaining years must absorb the steeper slope, and the options for absorbing it get more expensive the later they start. This is also the cadence regulators are converging on: under CSRD, ESRS E1 expects targets, progress against them and the underlying methods to be disclosed annually, which is the same close produced for a different audience (the mechanics are covered in our CSRD reporting guide). Treat the interim target the way finance treats a covenant: measured every period, on defined terms, with variances explained in writing before anyone outside the building asks.

Market-based instruments in a net zero ledger

Renewable energy certificates and power purchase agreements are the instruments most likely to confuse a net zero ledger, because they are legitimate and tightly bounded at the same time. Under the GHG Protocol's Scope 2 guidance a company reports two electricity figures: location-based, using the average grid factor for the region, and market-based, reflecting contractual instruments such as certificates, green tariffs and PPAs. Quality criteria apply to the market-based figure: the instrument must come from the same market as the consumption, carry a credible vintage, and be retired exclusively for the reporting company, with the retirement statement kept as evidence. A certificate that fails those tests does not move the number. The two-figure structure exists precisely so that readers can see how much of a reduction is contractual rather than physical, and both figures belong in the ledger (background in our Scope 1 and Scope 2 primer).

The bounded part matters more than the mechanics. Market instruments act on Scope 2 and nothing else, and the roughly 90% reduction expectation is measured against the whole footprint. For a typical company whose Scope 2 is 10 or 15% of the total, buying every certificate on the market still leaves the pathway substantially untouched: Scope 1 combustion and Scope 3 purchasing dominate, and no contractual instrument exists that papers over them. Within Scope 2 itself there is a quality gradient worth respecting: a PPA that finances new generating capacity is a stronger claim than unbundled certificates bought on the spot market, and disclosure frameworks increasingly ask which one you hold. The ledger treatment is an instrument register: contract terms, volumes, vintages and retirement evidence, held alongside the inventory rather than netted into it, so that the market-based figure can be rebuilt from documents on demand.

A worked example: restating the baseline after an acquisition

Take a 2,000-person manufacturer with a 2024 baseline of 120,000 tonnes CO2e and an interim target of 42% by 2030. In 2027 it acquires a smaller plant whose operations add roughly 12,000 tonnes a year, about 10% of the footprint. Without a restatement, the 2027 inventory jumps and the company appears to have moved backwards despite genuine reductions in its original operations; with a quiet, undocumented adjustment, nobody can reconcile the published history. The significance threshold in the company's base year recalculation policy, commonly set around 5% of the baseline, is what makes the decision mechanical rather than political: 10% exceeds the threshold, so the baseline is restated.

The mechanics run in one direction: the past is rewritten to look like the present. The acquired operations' emissions are estimated for 2024 using the same boundary and methods as the rest of the baseline, from the target's own records where they exist and from conservative estimates where they do not. The baseline becomes 132,000 tonnes. Every intervening year is restated the same way, so a reported 8% reduction in 2026 is recomputed on the restated series and may land somewhat higher or lower than first published. The target percentage does not change, but the absolute tonnes it implies do: 42% off 132,000 is a different quantity of tonnes than 42% off 120,000, and the capital plan behind the pathway has to absorb the difference. The whole event is logged as a restatement: date, trigger, method, before and after totals, and which previously published figures it supersedes. Organic growth, by design, triggers nothing, because absolute targets exist precisely to keep growth inside the commitment. What makes any of this survivable is the machinery from decision 1: a baseline that exists as a recomputable ledger rather than a number in a slide deck.

What this means operationally, this quarter

• Build the inventory before the pledge: complete Scope 1 and 2 from meters, screened Scope 3 from spend, methods labeled per line. The Scope Classifier demo shows that screening pass live.
• Freeze the baseline as a ledger, not a PDF: data, factors and methods versioned together.
• Pick the target year and pathway with the CFO in the room, because the reduction curve is a capital allocation plan wearing a sustainability badge.
• Set up the annual close: same boundary, same rigour, restatements explicit, factor updates isolated. The regimes that will audit this (CSRD's ESRS E1 wants targets and progress, ISSB wants transition plans) are summarised on scope 3 emissions reporting.
• Write the claims language last, from the ledger, and let legal read it against the numbers.

The uncomfortable summary: net zero accounting is 10% pledge and 90% bookkeeping, sustained for a decade or two. Companies fail at it the way they fail at any long accounting obligation, through staff turnover, method drift and lost lineage, which is why the foundational investment is not the target announcement but the ledger underneath it. Start with what carbon accounting actually is, get the inventory defensible, and the net zero program inherits its credibility from there.