Promised synergies need a ledger

The deck says US$1 billion of synergies. The later filing says about US$800 million of annual cost synergies by the end of 2025, against an approximately US$1 billion target by the end of 2026. The tempting calculation is 80 percent. The defensible answer is different: the later statement is a management-reported progress figure, the target date has moved, and the public record does not supply the bridge a percentage would need.

That is not a technicality. A deal promise and a later statement are different evidence objects. If the unit, baseline, timing, scope, or owner changes, the number may still be important, but it is not the result of the first number. The work is to keep those differences visible before a board turns a progress sentence into a score.

The operating discipline is close to a living business-case control loop: every number keeps an owner, a timing rule, and a next review.

A promise has a unit

The first field in a synergy ledger is not the amount. It is the kind of amount.

Transaction value is not a synergy. Purchase consideration is not a synergy. Sale proceeds are not a synergy. Operating contribution, free-cash-flow accretion, an integration cost, an EPS effect, and a run-rate cost reduction may all appear in the same announcement or filing, but they answer different questions. A ledger that drops them into one column has already decided that they are comparable.

Pfizer’s Seagen record shows why the time field belongs beside the amount. The 2023 announcement said “nearly”, not approximately: Pfizer expected “nearly 1 billion in cost efficiencies in the third full year after the completion of the transaction". Pfizer's 2025 Form 10-K later reported approximately US\800 million of annual cost synergies by the end of 2025 and an approximately US$1 billion target by the end of 2026. The later figure is useful. It is not an audited realization rate against the earlier sentence. Its source, timing, and target status need to travel with it.

That is also why the integration plan spends what the deal bought treats the case as a controlled sequence rather than a headline.

Chevron’s Hess record makes the same point with a revision. The transaction announcement described about US$1 billion of pre-tax run-rate cost synergies within a year of closing. The second-quarter 2026 results release later described US$1.5 billion of annual run-rate synergies related to Hess. This page previously said a 2025 Form 10-K had reported the initial target achieved; no such filing is cited or held here, so the sentence is removed rather than left resting on a document nobody can open. That is a management series with an upward revision. Without a category bridge or a common baseline, it cannot be written as 150 percent realization.

ConocoPhillips and Marathon Oil add a different boundary. The announcement described at least US$500 million of run-rate cost and capital savings in the first full year. The later filing reported more than US$1 billion of run-rate synergies and about US$1 billion of separate one-time benefits, primarily related to tax effects. The two later fields may both matter to the integration case. They are not one recurring synergy number.

Newmont and Newcrest offer the cleanest milestone and still require a label. Newmont announced US$500 million of annual pre-tax synergies within 24 months. Its third-quarter 2024 results release reported a US$500 million annual run-rate more than a year ahead of schedule and described the figure as a management estimate rather than a GAAP or non-GAAP measure. That supports a management-reported milestone. It does not create an independently reconstructed benefit.

Why is the M&A synergy panel twelve discrete transaction rows rather than an industry average?

To make the comparison inspectable, I froze a small public panel. A row had to meet all of these conditions: announcement in 2023 or 2024, at least one US or European participant, a headline value of at least US$10 billion in the primary transaction record, completion by 28 August 2026, and a held completion record plus later public report or regulated filing. The value screen is the selection rule. It is not a claim that transaction value predicts synergy.

The panel includes acquisitions, mergers, a control transfer, and a business carveout. It keeps enterprise value, equity value, consideration, proceeds, operating contribution, and synergy fields separate. Its states describe the public record, not the quality of management execution.

The panel does not answer how often large mergers succeed. Twelve is the number of rows a reader can inspect under this rule. It is not a market denominator. A percentage formed by dividing a count of states by twelve would be a new claim, and the public records do not authorize it.

How should integration committees track individual cost line items before aggregate rates?

The useful object is not a dashboard with a green cell. It is a row that makes a later calculation earn its right to exist.

The worksheet forces a question that a percentage hides: same as what? If the promise was an annual run-rate and the later statement is a three-year aggregate, the fields may belong in the same deal history but not in the same arithmetic. If the promise was cost and the later statement is revenue, the difference is not a rounding issue. If the target date moved, the date is part of the result.

What the later number does not say

The later record often contains a number that looks close enough. Three traps recur.

Operating contribution is not synergy realization. A filing can report revenue or operating income contributed by the acquired business. That tells you something about the business perimeter and the reporting period. It does not isolate the benefit created by combining two operations.

Integration cost is not the inverse of synergy. A cost to integrate, retain people, or separate a business may be necessary to capture a benefit. It belongs in the investment case. It should not be subtracted from a later operating field unless the model defines the relation and keeps the horizons consistent.

Silence is not zero. Extra Space’s later filing gives merger-associated revenue, expense, and transition-cost fields without a same-scope bridge to its US$100 million annual run-rate operating synergy target. Gallagher’s pro forma exhibit explicitly excludes expected cost savings, operating synergies, revenue enhancements, and the costs needed to achieve them. Those are useful negative disclosures. Neither says that the promise failed.

The same rule applies to the rows with no quantified result. BlackRock’s later HPS filing supplies operating and accounting measures. DSV’s later report supplies a buyer-side impact and a later target, while the seller’s transaction announcement supplied no matching synergy baseline. HPE’s later filing introduces a target that was not in the announcement. These are reporting boundaries, not blanks to be filled with an assumption.

Build the ledger before the deal closes

The ledger is useful before the first integration meeting because the original promise is easiest to lose after the close. I would record nine fields.

1. Deal and structure. Acquisition, merger, majority-control transfer, or carveout. Keep the reporting perimeter visible.
2. Promise document and date. Announcement, merger filing, investor presentation, or another primary record. Save the original currency and wording in the source record.
3. Unit and baseline. Cost, revenue, cash flow, capital, headcount, capacity, or another defined measure. Record whether the amount is annual, cumulative, run-rate, or one-time.
4. Dependency and owner. Procurement, systems, footprint, pricing, cross-sell, talent, or another mechanism. Name the team that can observe it.
5. Target date. A date turns a promise into a testable event. If the date moves, preserve the old date and add the new one.
6. Later document and date. Annual report, results release, proxy, or another public record. Do not let a later document inherit the first document’s label automatically.
7. Later field. Copy the later amount in its source unit. Keep operating contribution, accounting fields, integration cost, and synergy fields separate.
8. Same-scope test. Compare unit, baseline, currency, timing, perimeter, and category. Mark the comparison partial when one of them is missing.
9. Verdict. Use a small vocabulary: confirmed, revised, not quantified, not comparable, or unresolved. The vocabulary describes the record. It does not rate the deal team.

This process also protects a good result. ONEOK’s later release reports approximately US$160 million to US$190 million of combined first-year cost and initial commercial synergy impacts after its Magellan acquisition. That is a meaningful management-reported result. The announcement had named cost, operational, tax, and commercial categories without one quantified baseline, so the ledger should preserve the result and its comparison limit at the same time.

Why must the realization verdict remain unresolved until verified by post-close audits?

The pressure to publish a binary result is understandable. A board wants to know whether the thesis worked. But a public record can be precise about a different thing: what was promised, what was later said, and whether the two can be reconciled.

That is why the panel carries a state rather than a score. A management milestone can be worth acting on while still being management-reported. A revised number can be better information while still not being comparable with the first target. A filing can be silent because the company reports the measure elsewhere, because the target is not a required disclosure, or because the scope has changed. The ledger records that the bridge was not found. It does not invent the bridge.

For the board, the questions are short:

• What was the original unit and baseline?
• Has the target date passed, and did the perimeter stay the same?
• Is the later number recurring, aggregate, operating, accounting, or one-time?
• Is it a company statement, an audited line, or an inference?
• If the fields do not match, what decision is being made without the missing bridge?

The last question is the point. A ledger does not make a synergy real. It makes the decision-maker notice which part is real, which part is reported, and which part is still a promise.

Where are the empirical boundaries of M&A synergy accounting?

This is a public, receipt-bounded panel of twelve transactions announced in 2023 or 2024 and completed by 28 August 2026. It is not a census of M&A, a representative sample, or a synergy success rate. Management-reported milestones remain attributed as such. A later public filing that does not provide a same-scope bridge is coded as not quantified, not comparable, or unresolved, never as zero. The panel does not contain private company data, customer data, or a claim about any deal outside its named records.

Evidence base. The analytical frame also draws on these additional sources: Pfizer 2023; Pfizer 2026; Chevron 2023; Chevron 2026; ConocoPhillips 2024; Newmont 2023; ONEOK 2024; Extra Space Storage 2025; Hewlett Packard Enterprise 2025; DSV 2026; Carrier 2025; Arthur J. Gallagher & Co 2025; BlackRock 2026. The links identify the exact works; they support the mechanisms and boundary conditions discussed here, not every claim in isolation.