Skip to main content

Non-GAAP Adjustments

Updated Jun 24, 2026 at 2:35pm

Research Draft High 1,231 words

Non-GAAP adjustments are the line items a company adds back to, or subtracts from, its GAAP (Generally Accepted Accounting Principles) net income to produce a "tailored" earnings figure — usually labeled adjusted EPS, adjusted net income, adjusted EBITDA, or "core/operating" earnings. Management's argument is that GAAP earnings are distorted by one-time, non-cash, or non-operational noise, and that stripping it out reveals the firm's true recurring earning power. The core tension is that the same flexibility that legitimately improves comparability also lets management self-define profitability — and because management chooses what counts as "non-recurring," the adjustments almost always flatter results. Non-GAAP earnings are therefore simultaneously among the most useful and the most abused numbers in financial reporting, which is exactly why they belong in a forensic / red-flags toolkit.

How they're calculated / formed

A non-GAAP measure starts from a GAAP figure and applies named add-backs. The most common categories (per Audit Analytics and an S&P 100 study summarized in The CPA Journal, 2020) are:

  • Stock-based compensation (SBC) — the single most common exclusion, dominant in tech/SaaS. It is non-cash but recurring and real (it dilutes shareholders).
  • Amortization of acquired intangibles — non-cash, tied to past M&A.
  • Restructuring / severance charges — recurring at serial acquirers despite the "one-time" framing.
  • Impairments and write-downs — drove the largest single adjustments in the CPA Journal study.
  • Tax effects of the above, plus discrete tax items. Tax + restructuring together made up roughly 50% of total adjustments in that study.
  • Litigation settlements, M&A/transaction costs, FX, "non-recurring" items.

Adjusted EBITDA is the most aggressive common variant: it starts from net income and adds back interest, taxes, depreciation, amortization, and then a further layer of SBC, restructuring, and "one-offs." A reconciliation table from the non-GAAP figure back to the nearest GAAP measure is mandatory under SEC rules (Regulation G and Item 10(e) of Regulation S-K) in filings and earnings releases.

How they're used in practice

Analysts and "Street" consensus estimates are built on non-GAAP earnings far more often than GAAP — consensus EPS, P/E multiples, and earnings-surprise calculations typically run off the adjusted number. For a forensic analyst the value is not the adjusted figure itself but the bridge: read the reconciliation line by line and ask of each add-back:

1. Is it truly non-recurring? "Restructuring" that appears every year, or "one-time" acquisition costs at a roll-up that acquires every quarter, is recurring and should not be excluded. 2. Is it non-cash and economically immaterial to owners? SBC is non-cash but transfers value away from shareholders — excluding it overstates real profitability. 3. Is the treatment symmetric? Excluding non-recurring charges while keeping non-recurring gains is a classic abuse the SEC explicitly flags. 4. Has the definition stayed consistent? A measure that changes composition when GAAP earnings disappoint is a warning sign.

The most decision-useful single metric is the GAAP-to-non-GAAP gap and its trend. A small, stable, well-explained gap signals genuine clean-up; a large gap that widens precisely when GAAP results weaken signals window-dressing.

Adoption, debate & evidence

Adoption is near-universal. Audit Analytics has reported the share of S&P 500 firms using a non-GAAP metric rising to roughly 97% in 8-K/10-K filings by 2017 (commonly cited; ~88% in its earlier Q3-2015 earnings-release sample), and that non-GAAP adjustments increase reported income roughly 82% of the time (228 of 278 firms in its sample) — i.e., the bias is overwhelmingly upward, not random clean-up. The S&P 100 study (CPA Journal, 2010–2016) found 77% of adjustments were positive vs. 23% negative.

The academic evidence is genuinely two-sided and should not be smoothed over:

  • Pro-information camp: Multiple studies find non-GAAP earnings are more value-relevant and better predictors of future operating earnings than GAAP, and that this persists after Regulation G (2003) and the SEC's 2010 C&DI guidance. Recent work (Gies/Illinois, 2025) finds transparent non-GAAP disclosure makes M&A pricing more efficient.
  • Pro-skeptic camp: Equally credible research finds managers tailor non-GAAP earnings to meet or beat targets when GAAP falls short, and that aggressive exclusion of recurring items predicts future restatements — including fraud restatements. A widely cited working paper (Guest, Kothari & Pozen, Cornell/MIT, HBS-hosted, 2018) found high non-GAAP earnings predict abnormally high CEO pay, since incentive comp is often tied to the adjusted figure.

The honest synthesis: non-GAAP earnings carry real incremental information on average, but they are a conflicted disclosure whose bias is reliably optimistic and whose quality varies enormously by firm. They inform, and they manipulate — often in the same filing.

Strengths & limitations

When they help: removing genuinely transient noise (a one-time legal settlement, a single large impairment, purchase-accounting amortization) so two periods or two firms can be compared; surfacing cash-generative power obscured by heavy non-cash GAAP charges.

When they fail / the #1 misuse: recurring costs dressed as one-time exclusions — serial "restructuring," perpetual "integration costs," and especially SBC treated as if it weren't a real cost to owners. The SEC's most-cited prohibited practices are (a) excluding normal cash operating expenses to manufacture a measure ("individually tailored" recognition that changes a GAAP principle), and (b) presenting the non-GAAP figure with more prominence than GAAP. Buffett has repeatedly criticized SBC exclusion ("if compensation isn't an expense, what is it?"). Notably, some firms — e.g., Alphabet — have moved toward including SBC, which sophisticated investors read as a maturity / earnings-quality signal. SBC exclusion is also age-dependent: per Equity Methods, >90% of newly public firms exclude SBC vs. <25% of firms 20+ years old.

The deepest limitation is non-comparability: because each company defines its own "adjusted" measure, two firms' adjusted EPS are not on the same basis, and a single firm's definition can drift over time. Treat any non-GAAP number as a hypothesis to audit via the reconciliation, never as a fact.

Sources

Dispute flagged: the academic literature is genuinely split — value-relevance studies and manipulation/restatement studies both hold up. This doc treats non-GAAP earnings as informative-but-biased rather than endorsing either pole. Some specific percentages (e.g., 82% income-increasing, and prevalence figures of ~88% in earnings releases / ~97% in filings) are from Audit Analytics summaries of specific sample years and scopes, and will drift over time — the two prevalence numbers measure different documents and should not be conflated.