GeminIQ Website Logo
Subscribe

Adjusted EPS: Definition, Formula, and Why It Varies

Chad Hartman

By

Published · Last updated

Every other figure on an income statement has a rulebook behind it. Revenue recognition runs on ASC 606. Lease accounting runs on ASC 842. Earnings per share runs on ASC 260, which specifies the numerator, the denominator, and the treatment of every dilutive instrument in between. Adjusted EPS runs on nothing. No accounting standard defines it, no regulator specifies its inputs, and no two sources are obligated to calculate it the same way. It is the most widely quoted per-share figure in financial media that has no authoritative definition at all — which means the number carries information about whoever calculated it, not just about the company it describes.
Formula Adjusted EPS = (Net Income TTM + Stock Based Compensation TTM) ÷ Basic Shares Outstanding
Net Income (TTM) Trailing twelve-month GAAP net income, exactly as filed
Stock Based Compensation (TTM) Trailing twelve-month SBC expense, taken from the cash flow statement
Basic Shares Outstanding Period-end basic share count — not the diluted count
Typical gap vs. GAAP EPS Minimal for cash-compensation businesses; wide for software and semiconductors, where SBC commonly runs 10% to 30% of revenue
Not added back Restructuring charges, intangible amortization, impairments, litigation settlements, acquisition costs

Start your 7-day free trial →

Table of Contents

What Is Adjusted EPS?

Adjusted EPS is earnings per share recalculated after adding selected expenses back to net income. The logic behind it is narrow and legitimate: some expenses reduce reported earnings without reducing cash, and a per-share figure that strips those expenses out shows what the business generated in cash terms rather than accounting terms.

Stock-based compensation is the canonical case. A company that pays an engineer $200,000 in salary writes a check for $200,000. A company that pays the same engineer $120,000 in salary and $80,000 in restricted stock writes a check for $120,000 and issues equity for the rest. Both recorded $200,000 of compensation expense. Only one of them spent $200,000. Adjusted EPS exists to make those two companies comparable on a cash basis.

The problem is what happened to the concept afterward. The same reasoning that justifies adding back stock compensation has been extended to restructuring charges, acquisition costs, litigation settlements, intangible amortization, and in some cases to losses a company simply prefers not to discuss. Each extension arrives with the same defense — this expense does not reflect ongoing operations — and each extension moves the resulting figure further from anything a filing actually reports.

Why Is There No Standard Adjusted EPS Formula?

Because adjusted EPS is a non-GAAP measure, and non-GAAP measures are defined by whoever publishes them.

Regulation G requires a company presenting a non-GAAP figure to reconcile it to the nearest GAAP equivalent and to avoid presenting it more prominently than the GAAP number. That is a disclosure rule, not a definition rule. It governs how the adjustment must be shown. It says nothing about which adjustments are permitted, and it applies to the company's own reporting rather than to third-party platforms recalculating the figure independently.

So "adjusted EPS" describes a category of calculation rather than a specific one. A company's investor relations deck may add back four items. A sell-side model may add back two of those four plus one the company never mentioned. A data platform may add back one. Three different figures, three defensible methodologies, one company, one quarter. A reader comparing an adjusted EPS figure from one source against an adjusted EPS figure from another is frequently comparing two unrelated calculations that happen to share a label.

The same structural problem runs through normalized financial data generally, and it is worse here than almost anywhere else, because the adjustment is discretionary by design rather than by accident.

The Adjustments Companies Actually Make

Four adjustments account for most of the distance between GAAP EPS and a company-reported adjusted EPS figure, and they are not equally defensible.

Stock-based compensation is the most common and the most contested. It is non-cash in every sense that matters to the cash flow statement. It is also a real cost to existing shareholders, because the equity issued to employees dilutes everyone else's claim on the business. Adding it back produces a cleaner picture of cash generation and a distorted picture of shareholder economics at the same time.

Intangible amortization is the second, typically defended on the grounds that amortizing acquired customer relationships or developed technology reflects a purchase price allocation rather than an ongoing operating cost. That argument holds for a company that acquires rarely. It weakens considerably for a serial acquirer, where a steady stream of amortization is arguably the cost of the growth strategy itself.

Restructuring charges are the third, and the label does real work here. A single restructuring in a decade is plausibly non-recurring. Restructuring charges appearing in eleven consecutive quarters are an operating expense wearing a different name, and a company adding them back every quarter is reporting an earnings figure that excludes a permanent cost of its business. But standard financial media doesn't read the footnotes, and a charge described as one-time in a press release is rarely checked against the eight quarters that preceded it.

Litigation settlements, impairments, and acquisition costs make up the fourth group, and they share a pattern worth watching: they are almost always additions to earnings, rarely subtractions. Adjustment sets that move in one direction across every period are describing a preference, not a methodology.

How to Calculate Adjusted EPS

GeminIQ resolves the ambiguity above by fixing the adjustment set at exactly one item and applying it identically to every company in the database. Three steps, three inputs, all of them traceable to a filed statement.

Step 1: Start With Trailing Twelve-Month Net Income

The trailing-twelve-month basis matters more than it appears to. A single quarter's net income reflects whatever seasonality, one-time charge, or tax event landed in that quarter, and a per-share figure built on it inherits all of that noise. Summing four quarters smooths the seasonal pattern and puts the numerator on the same footing as the annual figures most benchmarks assume.

Net income itself requires no adjustment at this stage. It is the GAAP figure exactly as the company filed it, summed across the trailing four quarters.

When it breaks: A company that changed its fiscal year, completed a major divestiture, or restated a prior period has a trailing-twelve-month net income figure spanning two structurally different businesses. The arithmetic still works. The comparison to the prior year does not.

Step 2: Add Back Stock Based Compensation

Stock-based compensation appears in the operating activities section of the cash flow statement, where it sits as a non-cash add-back reconciling net income to operating cash flow. That placement is the tell: the company itself has already classified the expense as non-cash for cash flow purposes. Adjusted EPS applies the same classification to the earnings line.

Pull the figure on the same trailing twelve-month basis as net income and add it to the numerator. GeminIQ sources its pre-calculated Stock Based Compensation figure directly from the as-filed cash flow statement, with the tag attached, rather than reconstructing it from a compensation footnote or a normalized feed. GeminIQ publishes the intermediate result — net income plus SBC, before any per-share division — separately as Adjusted Net Income.

When it breaks: Some companies capitalize a portion of stock compensation into software development costs or inventory rather than expensing it in the period. The capitalized portion never appears as an expense on the income statement, so it is not in net income and cannot be added back to it. The cash flow statement's SBC line generally captures total expense recognized, not total SBC granted, and the gap between those two figures stays invisible without reading the equity footnote.

Step 3: Divide by Basic Shares Outstanding

The denominator is the period-end basic share count, not the diluted count. That choice is deliberate, and it is the single most common source of confusion when a by-hand calculation fails to reconcile.

Basic shares keep the metric on the same basis as market capitalization and every other per-share ratio built from it, which makes adjusted EPS directly comparable against Free Cash Flow Per Share without a denominator mismatch. Using the diluted count instead would set a cash-basis numerator against a share count that already anticipates the dilution the numerator just added back — counting the same equity from both directions.

When it breaks: For a company issuing equity aggressively, a period-end share count understates the average shares outstanding across the period, and the resulting adjusted EPS runs slightly flattering. GeminIQ's Dilution Ratio shows how fast the count is moving, which is the context that tells you whether the point-in-time denominator is a reasonable proxy or a stale one.

What GeminIQ Adds Back — and What It Deliberately Doesn't

GeminIQ's Adjusted EPS adds back stock-based compensation and nothing else. No restructuring charges. No intangible amortization. No impairments, litigation settlements, or acquisition costs.

That constraint is the point. An adjustment set that varies by company cannot support a comparison between companies, because any difference in the resulting figures reflects a difference in adjustment policy as much as a difference in business performance. Fixing the adjustment at one item — the one with the clearest non-cash justification and the most consistent disclosure across filers — produces a figure that means the same thing everywhere.

The tradeoff is real and worth stating plainly. A company that took a truly non-recurring $2 Billion legal settlement will show a depressed GeminIQ adjusted EPS that its own investor deck would present very differently. GeminIQ's figure is not attempting to represent management's view of normalized earnings. It is attempting to be the same calculation for every company in the universe, which is a different property, and a more useful one when the goal is comparison rather than narrative.

Adjusted EPS vs. GAAP EPS vs. Diluted EPS

Three per-share figures describe the same company in the same period, and the distance between them is more informative than any one of them alone.

Diluted Earnings Per Share is the conservative floor. It divides GAAP net income by a share count that includes every outstanding option, restricted share, and convertible instrument as though all of them converted. It answers the least flattering version of the question: if every equity claim on this business were exercised today, what would each share have earned?

Basic GAAP EPS sits above it, dividing the same net income by the actual current share count and ignoring the instruments that have not converted yet.

Adjusted EPS sits highest, because it has added an expense back to the numerator that the other two subtract.

The spread carries the signal. A business paying its workforce primarily in cash produces three figures clustered tightly together, because there is little SBC to add back and few unexercised instruments to dilute against. A software or semiconductor business where stock compensation runs 10% to 30% of revenue produces a wide spread — and that spread quantifies how much of the reported earnings stream is routed to employees as equity rather than retained for shareholders. GeminIQ's Stock Based Compensation To Revenue measures the same phenomenon directly.

Reading adjusted EPS on its own discards that information entirely. Reading it against the diluted figure recovers it.

When Adjusted EPS Breaks

The metric holds up as a comparison tool across most businesses. Three situations distort it enough to warrant caution.

A company with negative net income can produce a positive adjusted EPS purely from the add-back. A business losing $400 Million with $600 Million in stock compensation shows $200 Million in adjusted net income and a positive adjusted EPS while the GAAP figure stays negative. Neither number is wrong. Presenting the positive one without the negative one is where the distortion enters, and this pattern shows up precisely where SBC is largest — early-stage software companies with heavy equity compensation and no GAAP profitability.

Rapid share issuance breaks the denominator's reliability. Adjusted EPS can rise year over year while total adjusted net income falls, or fall while the business improves, purely on share count movement. Any per-share figure needs the share count trend beside it, and stock compensation is itself one of the primary drivers of that trend — the mechanism traced in detail in Stock Buybacks vs. Stock Compensation.

Cross-source comparison breaks by default rather than by exception. An adjusted EPS figure quoted in a headline, an earnings deck, and a data platform are three different calculations unless someone has verified the adjustment sets match. When two sources disagree, the disagreement is almost always methodological rather than factual. Identify which adjustments each one applied and the discrepancy resolves faster than any recheck of the arithmetic.

Frequently Asked Questions

What is adjusted EPS in simple terms?

Adjusted EPS is earnings per share calculated after adding certain expenses back to net income, most commonly stock-based compensation. It shows what per-share earnings would look like if those expenses were excluded. No accounting standard defines which expenses qualify, so the specific adjustments vary by whoever publishes the figure.

Is adjusted EPS the same as non-GAAP EPS?

They overlap but are not interchangeable. "Non-GAAP EPS" is the broader regulatory category covering any per-share figure that departs from GAAP. "Adjusted EPS" is a common label applied to figures inside that category. A company's own non-GAAP EPS typically carries several adjustments chosen by management, while GeminIQ's Adjusted EPS applies exactly one — the stock-based compensation add-back — to every company identically.

Why do different websites show different adjusted EPS for the same company?

Because they are adding back different things. One source may add back stock compensation only, another may add stock compensation plus intangible amortization plus restructuring charges. Both figures can be internally consistent and still disagree by a wide margin. Checking which adjustments each source applies resolves nearly every discrepancy of this kind.

Should I use adjusted EPS or diluted EPS?

Both, read together. Diluted EPS gives the conservative view using a fully diluted share count. Adjusted EPS gives the cash-basis view before a major non-cash charge. The gap between them measures how much of the earnings stream stock compensation is consuming, and that gap is more informative than either figure taken alone.

Where does the stock-based compensation figure come from in a filing?

The operating activities section of the cash flow statement, where SBC appears as a non-cash add-back reconciling net income to operating cash flow. Pulling it from GeminIQ's Financial Statements view returns the as-filed figure with its XBRL tag attached, rather than a value reconstructed from a compensation footnote or reclassified by a third-party feed. The equity footnote carries additional detail on grants and vesting that the cash flow line does not.

Can adjusted EPS be positive when the company is losing money?

Yes, and it happens routinely. If stock-based compensation exceeds the net loss, adding it back flips the numerator positive while GAAP net income remains negative. The calculation is valid. Reading the adjusted figure without the GAAP figure alongside it is not.

Start your 7-day free trial →

Wall Street's data. Main Street's price.

Institutional terminals charge thousands a year for as-filed accuracy. GeminIQ gives you the same thing for a fraction of the cost: financials built directly from raw SEC EDGAR filings, not third-party APIs, with full XBRL traceability back to the original 10-K or 10-Q. No normalized guesswork, just calculated metrics, charts, screeners, and watchlists built on numbers exactly as the company reported them. Start researching now at GeminIQ.com.

Disclaimer: The content in this blog is for educational and entertainment purposes only and does not constitute financial, legal, or tax advice. Investing involves risk, including the loss of principal. The views expressed are my own and not intended as financial advice or a guarantee of future performance.