EBITDA Add-Backs: Which Adjustments Hold Up in Diligence

Why add-backs decide the price

Many lower-middle-market offers are written as a multiple of an earnings figure, and the figure is rarely reported EBITDA. It is the seller's adjusted EBITDA: reported EBITDA plus a column of "add-backs" that the seller argues do not reflect what the business will earn under a new owner. The arithmetic makes every line in that column expensive. If the offer is a multiple of adjusted EBITDA, each dollar of accepted add-back is worth the multiple in purchase price, and each dollar the buyer accepts in error is paid for at the same rate.

A buyer can agree a multiple in an indication of interest and still lose the economics in diligence if the bridge from reported to adjusted EBITDA is accepted without evidence. This article describes practice and public rules, including the U.S. Securities and Exchange Commission's non-GAAP tests; it is not accounting, tax or legal advice.

The four families of adjustment

Nearly every bridge line falls into one of four families; classify each line first, because the family tells you which evidence to ask for.

The families sit on a gradient of evidence: owner items can usually be proven from payroll and the ledger, one-time items can be tested against history, run-rate items need contracts plus early actuals, and pro forma items are forecasts. The further down the table, the more an add-back is a projection being priced as a fact.

Two points sharpen the table. The valid owner adjustment is a normalization, not a deletion: if the owner drew $450,000 and a general manager would cost $220,000 fully burdened, the add-back is the difference, and it shrinks further if the owner also performs a second role (estimator, head of sales, key account manager). Run-rate items should be stated as contribution margin, not revenue, and should already be visible in invoices.

What the SEC treats as misleading, and why it is a useful yardstick

A boundary first: the SEC's non-GAAP rules govern disclosures by SEC registrants. Regulation G applies when "a registrant, or person acting on its behalf" publicly discloses a non-GAAP measure, and Item 10(e) of Regulation S-K applies to measures in SEC filings. A private seller's CIM is not governed by either. For a private buyer they are the most developed public statement of the tests an adjusted figure should pass, enforced by a regulator: a yardstick, not law.

Five tests carry over cleanly:

1. Normal, recurring, cash operating expenses. The SEC staff's interpretation 100.01 says that presenting a performance measure that "excludes normal, recurring, cash operating expenses necessary to operate a registrant's business is one example of a measure that could be misleading." The same answer says the staff views an operating expense "that occurs repeatedly or occasionally, including at irregular intervals, as recurring" (SEC staff non-GAAP interpretations, Q100.01, updated December 13, 2022). Irregular is not non-recurring.
2. Charges without gains. A measure "adjusted only for non-recurring charges when there were non-recurring gains that occurred during the same period" could violate Regulation G (Q100.03). A bridge that adds back the lawsuit but leaves the gain on the equipment sale in EBITDA fails this test.
3. Consistency between periods. Adjusting a charge in the current period while similar charges in prior periods were not adjusted can be misleading unless the change is disclosed and explained (Q100.02). Ask for the bridge for every year presented, not only the latest.
4. The two-year recurrence rule. Item 10(e)(1)(ii)(B) prohibits adjusting a performance measure "to eliminate or smooth items identified as non-recurring, infrequent or unusual, when the nature of the charge or gain is such that it is reasonably likely to recur within two years or there was a similar charge or gain within the prior two years" ^[1]. Staff interpretation 102.03 clarifies that the prohibition governs the description: a registrant may still make an adjustment it considers appropriate, but cannot call it non-recurring unless it meets the criteria (Q102.03). For a buyer: look back two years for a similar item before accepting the word "one-time".
5. Recognition and measurement. Adjustments that change GAAP's recognition pattern, for example accelerating ratably recognized revenue as if earned when billed or switching from accrual to cash basis, are given as examples of measures that may be misleading (Q100.04). Small-company bridges that mix cash-basis books with accrual adjustments should be read with this in mind.

Two further points apply directly to CIMs. A measure labelled "pro forma" that is not calculated consistently with Article 11 of Regulation S-X is mislabelled (Q100.05), so "pro forma EBITDA" in a CIM must be defined line by line. And disclosure does not cure everything: in the staff's view a measure can mislead to a degree that even "extensive, detailed disclosure about the nature and effect of each adjustment would not prevent" it from being materially misleading (Q100.06).

Regulation G also requires a quantitative reconciliation to the most directly comparable GAAP measure ^[2]. The private-deal equivalent is a bridge that starts from net income on the financial statements, not from an EBITDA figure the seller computed.

Enforcement examples: prominence and misclassification

Two settled SEC actions against public companies show two ways adjusted-earnings presentations go wrong.

Prominence (ADT, 2018). Item 10(e)(1)(i)(A) requires the most directly comparable GAAP measure to be presented with equal or greater prominence ^[1]. In its order of December 26, 2018, the SEC found that ADT's fiscal 2017 earnings release headline stated that adjusted EBITDA was up 8% year-over-year without mentioning net income or loss, and that the first-quarter 2018 release headline said adjusted EBITDA was up 7%, while the GAAP net loss appeared further down in the text ^[3]. ADT consented to a cease-and-desist order and a $100,000 civil penalty. For a buyer reading a CIM: when the adjusted figure is the headline and reported earnings sit in an appendix, start from the appendix.

Misclassification (DXC, 2023). The SEC's March 14, 2023 release states that DXC "materially increased its reported non-GAAP net income by negligently misclassifying tens of millions of dollars of expenses" as transaction, separation and integration-related costs, and that its controls were not adequate to keep classifications consistent with the company's own public description of those costs. Without admitting or denying the findings, DXC consented to an $8 million penalty and to develop non-GAAP policies and disclosure controls ^[4]. The lesson: an add-back category can be legitimate in principle and still be filled with ordinary operating costs in practice. Test the contents of the category, not just its name.

Evidence each add-back needs

Each common add-back, the evidence that makes it underwritable, and the red flags that should trigger a haircut or rejection:

The last row deserves emphasis because sellers rarely list it. Apply the SEC staff's charges-and-gains logic (Q100.03) and ask the seller to identify every non-operating or unusual gain in the period: asset sales, insurance recoveries, PPP or other grant income, one-off rebates.

Illustrative bridge (fictional company)

Illustrative. Brightwater Fluid Controls is fictional. Its reported FY2025 EBITDA is $2,400,000. The seller's bridge adds eight items to reach $3,610,000. The buyer tests each.

The gap is $880,000. At an illustrative multiple of 6.0x (an assumption for the arithmetic, not a market figure), that is $5.28 million of price. Note what the buyer did not do: it did not reject the run-rate items as false. It declined to pay for them at close and kept them available for structure, for example an earnout or a price adjustment tied to invoiced results, which is where an unproven but plausible item belongs.

Worksheet: add-back test sheet

Illustrative template. One row per add-back, filled in before any figure enters the valuation model.

Three rules: no add-back without an evidence reference (a blank cell means "pending", not "accept"); dispositions line by line, never a blanket percentage haircut that hides which claims failed; and the same IDs carried into the quality-of-earnings scope and post-close tracking of run-rate items.

What academic work says about non-GAAP aggressiveness

The research literature on non-GAAP reporting concerns public companies, but it frames the same question a buyer faces: are adjusted figures informative, or are they used to flatter results? Black, Christensen, Ciesielski and Whipple's synthesis in the Journal of Business Finance & Accounting reviews nearly two decades of that research. Its abstract notes that the number of firms reporting non-GAAP earnings "has increased dramatically over the last decade" and that the SEC, FASB and IASB have all questioned "whether these measures are misleading to investors"; the authors combine the literature review with a new dataset of detailed non-GAAP disclosures ^[5]. The full paper is paywalled, so its detailed findings are not summarized here; only the abstract, also indexed on RePEc/IDEAS, was reviewed.

For a private buyer the takeaway is the framing, not a number: adjusted earnings can be informative or can be shaped by the presenter's incentives, and the test sheet tells the two apart one line at a time.

Where public research stops and diligence starts

Most add-backs cannot be tested from outside the company: payroll, general-ledger detail, settlements and invoices arrive only after an NDA. Before that point the work is narrower:

• Before first contact: confirm the legal entity and its history, look for public litigation that bears on "one-time" legal costs, and check whether earlier events (a relocation, a prior lawsuit, a change of ownership) appear in public records within the two-year look-back window.
• On receipt of the CIM: classify every bridge line into the four families, flag any line without an amount and period, and list the evidence each line will need. Ask for the bridge for every year presented.
• In confirmatory diligence: hand the test sheet to the quality of earnings provider as part of their scope, so their work tests the seller's specific claims rather than producing a separate adjusted figure that cannot be reconciled to yours.

Two habits keep the handoff clean. Record each fact with its source and date, under the same kind of evidence standards you would apply to any other research claim. And pass unresolved items forward as named open questions rather than silent assumptions; the research handoff checklist shows one way to structure that packet.

Docket's Triage, Scout and Audit agents turn a firm's own screening criteria into questions answered from retained, quoted public sources, which covers the pre-NDA part of this work; the add-back tests themselves belong to the deal team and its accountants.

Sources

1. Non-GAAP Financial Measures, Corporation Finance staff interpretations (Q100.01-100.06, 102.03, 102.10) — U.S. Securities and Exchange Commission (last updated 2022), https://www.sec.gov/divisions/corpfin/guidance/nongaapinterp.htm, accessed 2026-10-28
2. 17 CFR § 244.100, General rules regarding disclosure of non-GAAP financial measures (Regulation G) — Legal Information Institute, Cornell Law School (current eCFR text), https://www.law.cornell.edu/cfr/text/17/244.100, accessed 2026-10-28
3. 17 CFR § 229.10, Regulation S-K Item 10 (paragraph (e)) — Legal Information Institute, Cornell Law School (current eCFR text), https://www.law.cornell.edu/cfr/text/17/229.10, accessed 2026-10-28
4. In the Matter of ADT Inc., Order Instituting Cease-and-Desist Proceedings, Release No. 34-84956 — U.S. Securities and Exchange Commission (2018), https://www.sec.gov/litigation/admin/2018/34-84956.pdf, accessed 2026-10-28
5. SEC Charges IT Services Provider DXC Technology Co. for Misleading Non-GAAP Disclosures, press release 2023-49 — U.S. Securities and Exchange Commission (2023), https://www.sec.gov/newsroom/press-releases/2023-49, accessed 2026-10-28
6. Black, D. E., Christensen, T. E., Ciesielski, J. T. and Whipple, B. C., "Non-GAAP reporting: Evidence from academia and current practice," Journal of Business Finance & Accounting 45(3-4): 259-294 — Wiley (2018), https://doi.org/10.1111/jbfa.12298, accessed 2026-10-28