How to Compare Adjusted Numbers
How to Compare Adjusted Numbers
Why Comparing Adjusted Numbers Matters
Comparing adjusted metrics across companies and time periods is one of the most powerful tools for detecting when accounting is being used to obscure economic reality. When all companies are adjusting earnings for different items in different ways, the question becomes: which adjusted metrics are meaningful, and which are marketing spin?
The problem is straightforward: there is no single "correct" definition of adjusted earnings. A software company might adjust for stock-based compensation, amortization of intangibles, and restructuring costs. A manufacturer might adjust for facility closures, asset write-downs, and pension-related gains. A bank might adjust for goodwill impairments, securities losses, and regulatory settlements. Each company's definition is defensible, but they're not directly comparable—making peer comparison and trend analysis devilishly difficult.
The antidote is methodical comparison: examining what each company excludes, normalizing those exclusions for consistency, and asking whether the adjusted metrics reveal real performance differences or accounting differences. This work requires more time than trusting management's adjusted numbers, but it prevents being misled by companies whose adjusted earnings have divorced from economic reality.
Quick definition: Comparing adjusted numbers means examining what each company adjusts for, normalizing definitions across peers, and verifying that reported adjustments reflect real economic events rather than accounting engineering.
Key Takeaways
- There's no universal definition of adjusted metrics; companies define adjustments based on their own business model and accounting choices
- Peer comparison requires normalizing adjustments so metrics are truly comparable across companies
- Multi-year trend analysis reveals whether adjustments are truly one-time or have become recurring
- Cash flow comparison exposes when adjusted earnings growth isn't supported by real cash generation
- Industry-specific metrics (like EV/EBITDA for capital-intensive businesses) require understanding typical adjustments in that sector
- Suspicion should increase when a company's adjusted metrics diverge significantly from both peers and GAAP results
Building a Normalized Earnings Model
The first step in comparing adjusted numbers is building your own normalized earnings model. Rather than accepting each company's definition of adjusted earnings, start with GAAP net income and apply consistent adjustments across all companies you're analyzing.
For example, if you're comparing three software companies:
- Company A excludes stock-based compensation, amortization of intangibles, and restructuring costs from adjusted earnings.
- Company B excludes stock-based compensation and amortization of intangibles but includes restructuring costs.
- Company C excludes only amortization of intangibles.
Their reported adjusted earnings figures are not directly comparable. To make them comparable, you would:
- Start with each company's GAAP net income.
- Add back (or exclude) the same items for all three companies in the same way.
- Arrive at a normalized earnings figure for each.
This normalized figure is your own adjusted metric, derived from consistent rules. It might differ from what management reports, but it's comparable across companies.
For example, a normalized metric might be: GAAP net income + stock-based compensation + amortization of intangibles + one-time restructuring charges (defined as charges exceeding 1% of revenues). Using this definition consistently allows you to compare the three software companies on equal footing.
Identifying What Companies Adjust For
Before normalizing, you need to identify what each company adjusts for. This information appears in the reconciliation of non-GAAP to GAAP metrics, typically in the earnings release or 10-Q/10-K filing.
Common adjustments:
- Stock-based compensation: Nearly all technology companies exclude this. It's a real cost (dilutive to shareholders) but also non-cash and can vary significantly with stock price. Comparing software companies almost always requires normalizing for stock-based compensation.
- Amortization of intangibles: Companies acquired by larger firms, or companies that have made acquisitions, exclude this from adjusted metrics. The adjustment is defensible (amortization is a non-cash charge), but it's also real: the company paid for those intangibles and they do decline in value.
- Restructuring charges: These are highly variable and often claimed to be non-recurring, but savvy investors verify whether "restructuring" has become routine.
- Acquisition-related costs: Including integration costs, severance, and contract termination fees related to mergers and acquisitions.
- Goodwill impairment: When a company writes down the value of an acquisition it overpaid for, this appears as a one-time charge.
- Asset sales and gains/losses: Real transactions but often excluded as non-recurring.
- Tax-related items: Deferred tax benefits, tax rate adjustments, and one-time tax settlements are sometimes excluded.
Peer Comparison Framework
Once you've identified what each company adjusts for, comparison requires a systematic framework:
- Collect reported adjusted earnings for each peer over 3–5 years.
- Identify adjustments by reviewing reconciliations.
- Apply uniform adjustments to arrive at normalized earnings.
- Compare growth rates and margins across peers using normalized figures.
- Investigate divergences: If one company's normalized earnings growth diverges from peers, ask why.
For example, comparing retail companies on adjusted EBITDA: all retailers face similar cost structures (real estate leases, labor, inventory management). If one retailer's adjusted EBITDA margin is significantly higher than peers, either:
- Its business model is more efficient (genuinely superior management).
- It's adjusting more aggressively, excluding items other retailers include.
- It's in a different segment (e.g., e-commerce vs. brick-and-mortar).
Investigating the cause separates real competitive advantage from accounting engineering.
Multi-Year Trend Analysis
One-time items should appear once. Recurring items appear every year. By tracking adjusted earnings over multiple years and the items that change from year to year, you can identify whether a company is being honest about what's one-time and what's recurring.
Create a table like this for a company you're analyzing:
| Year | GAAP EPS | Adjustments | Adjusted EPS |
|---|---|---|---|
| 2019 | $1.50 | $0.50 | $2.00 |
| 2020 | $1.75 | $0.45 | $2.20 |
| 2021 | $1.20 | $0.70 | $1.90 |
| 2022 | $0.90 | $0.85 | $1.75 |
| 2023 | $0.75 | $1.10 | $1.85 |
In this example, GAAP earnings are declining sharply (down 50% over five years), but adjusted earnings are relatively stable. The gap between GAAP and adjusted earnings is widening, and adjustments have grown from 25% to 60% of reported earnings.
Questions to ask:
- Why are adjustments growing?
- Are the same items being excluded every year (suggesting they're recurring)?
- Is management adjusting away a genuine decline in profitability?
- Would a new investor in 2019 have anticipated this deterioration based on adjusted metrics?
A company with growing adjustments and deteriorating GAAP earnings is a red flag.
Cash Flow Verification
Adjusted earnings can be manipulated; cash flow cannot. Operating cash flow represents real money moving through the company. If adjusted earnings are growing but operating cash flow is flat or declining, the adjusted earnings are suspect.
Create a comparison:
| Year | Adjusted EPS | Operating Cash Flow/Share |
|---|---|---|
| 2019 | $2.00 | $2.10 |
| 2020 | $2.20 | $2.25 |
| 2021 | $1.90 | $1.70 |
| 2022 | $1.75 | $1.40 |
| 2023 | $1.85 | $1.10 |
In this example, adjusted earnings have held relatively steady while cash flow has declined 50%. This divergence indicates that adjusted earnings are inflated by accounting adjustments, not supported by actual cash generation. A company with this pattern is not a good investment, regardless of what adjusted metrics claim.
Comparing Across Industries
Some adjustments are industry-specific and necessary for meaningful comparison. Understanding these prevents false comparisons.
Software/Technology: Stock-based compensation is universal and typically large (15–25% of earnings). Comparing tech companies by GAAP earnings is meaningless because stock-based comp varies with stock price volatility, not business performance. All tech comparisons require normalizing for stock-based comp.
Financial Services: Banks and insurers face regulatory settlements, reserve changes, and goodwill impairments as normal course of business. Comparing banks on GAAP net income alone misses the persistent nature of these items. A normalized metric might exclude the most volatile items while including settlement costs as recurring.
Manufacturing: Capital-intensive manufacturers face asset write-downs and facility closures related to cyclical downturns. Comparing manufacturers on adjusted EBITDA is appropriate, as these charges are often legitimately non-recurring, but investors should verify that excluded items aren't signs of structural competitive decline.
Real Estate: REITs and real estate companies often adjust for mark-to-market gains/losses on investment properties and derivatives. These adjustments can be large and volatile. Comparing REITs requires understanding which adjustments are real economic gains/losses and which are accounting artifacts.
Comparative Adjusted Metrics Flowchart
Real-World Examples
Amazon vs. Competitors (2010–2015). Amazon reported GAAP losses or minimal profits while competitors reported significant net income. Investors who only looked at GAAP earnings thought Amazon was unprofitable; Amazon was actually investing heavily in growth (warehouses, infrastructure) and taking the costs immediately. Competitors, investing less, reported higher GAAP profits. Normalizing for capital intensity—adjusting for the different depreciation and amortization schedules—revealed that Amazon's economic profitability was more comparable to peers than GAAP numbers suggested. Once depreciation and amortization converged with capital spending, GAAP profits aligned with adjusted metrics.
General Electric (2000–2016). GE reported large adjusted earnings from GE Capital (its finance division), claiming that adjusted metrics separated the finance business from the industrial business. Competitors like Siemens reported GAAP profits without the financial engineering. Normalizing both companies' earnings by excluding financial services businesses revealed that GE's industrial performance was weaker than adjusted metrics suggested. When the 2008 financial crisis hit and GE Capital imploded, it became obvious that the adjusted metrics had masked structural weakness in the core business.
Theranos (2013–2015). Theranos reported adjusted revenue metrics that excluded products not yet commercialized, adjusted gross margins by excluding certain costs, and adjusted operating metrics to exclude R&D. The company's reported adjusted metrics looked impressive. Comparing to real diagnostics companies (Quest, Labcorp) revealed that Theranos's metrics were defined so differently they weren't comparable. The company had no real revenues, was adjusting away nearly all costs, and was misleading investors. The adjusted metrics were a cover for fraud.
Common Mistakes
Mistake 1: Using Different Adjusted Metrics for Different Companies Comparing one company's reported adjusted EBITDA to a peer's reported adjusted operating income is meaningless. Start with consistent definitions across all peers. Either use each company's GAAP metrics (comparable by definition) or develop your own normalized adjustment rules that apply uniformly.
Mistake 2: Ignoring the Size of Adjustments Relative to Net Income If a company's adjustments exceed net income (more common than investors realize), the adjusted metric is unreliable. A company with GAAP net income of $100 million adjusting away $150 million in charges to arrive at adjusted earnings of $250 million is making adjustments so large that the metric has little credibility. Trust narrows dramatically when adjustments are >25% of reported net income.
Mistake 3: Assuming Peer Margins Should Be Identical Different companies have different business models, cost structures, and strategies. A premium brand will have higher margins than a discount brand; a asset-light software company will have higher margins than a capital-intensive manufacturer. Comparing margins across different business models is misleading. Compare only to true competitors with similar models.
Mistake 4: Not Verifying Adjustments Against Cash Flow The ultimate test of whether adjusted earnings are real is whether they convert to cash. Companies with widening gaps between adjusted earnings and operating cash flow are engaging in earnings quality deterioration. Verification is essential.
FAQ
Q: If I adjust earnings myself, won't I miss management's insights? A: You won't miss insights; you'll gain consistency. Management's adjustments are useful data points, but they're defined for management's benefit, not yours. By applying uniform rules, you see the actual economics without management's spin. You can always compare your normalized metrics to management's reported metrics to understand their perspective.
Q: Is it fair to compare companies with different accounting methods? A: Absolutely—that's why normalization exists. GAAP allows different companies to use different depreciation schedules, inventory valuation methods, and revenue recognition approaches even within the same industry. Normalizing accounts for these differences so metrics are truly comparable. Without normalization, accounting method becomes a source of competitive advantage, which distorts analysis.
Q: What if companies in an industry all adjust for the same items the same way? A: Then the adjustments are standard practice and normalizing becomes straightforward. For example, all technology companies exclude stock-based compensation from adjusted metrics. In this case, you can use reported adjusted metrics directly for peer comparison, as they're already normalized by industry practice. But verify by reading the reconciliations—exceptions do occur.
Q: How many years of history do I need to identify trending adjustments? A: Three years is minimum; five is better. With only one or two years, a genuinely one-time item can appear recurring. With five years of history, patterns emerge clearly. If the same items are being excluded repeatedly, they're recurring, regardless of how management labels them.
Q: Should I include or exclude non-recurring items in normalized earnings? A: Exclude them consistently. Your normalized metric should represent sustainable earnings from core operations. One-time items—losses on asset sales, unusual litigation settlements, gains on pension remeasurement—should be separately identified and excluded. This allows trend analysis without distortion from noise.
Q: What if a company doesn't provide a reconciliation? A: This is a red flag. Public companies are required by Regulation G to provide reconciliation of non-GAAP metrics to GAAP. If a company presents adjusted metrics without reconciliation, it's either violating SEC rules or not actually presenting non-GAAP metrics (just using different terminology for standard calculations). Either way, contact investor relations and demand clarification.
Related Concepts
- Normalized Earnings: Earnings adjusted consistently using uniform rules across all companies being compared. Normalized earnings are more comparable than reported (GAAP or reported adjusted) earnings because accounting differences are controlled for.
- Quality of Earnings: How much reported earnings are supported by cash flow and core operations, versus being inflated by accounting choices or one-time items. High-quality earnings are consistent, repeatable, and convert reliably to cash. Low-quality earnings are volatile, adjusted for large items, and diverge from cash flow.
- Peer Benchmarking: Using the average or median metrics of peer companies as a benchmark for evaluating a company's performance. If a company's margins are 300 basis points above peers, either its business model is superior or its accounting is more aggressive. Benchmarking reveals which is true.
- Earnings Sustainability: The degree to which current earnings are likely to persist into future periods. Earnings driven by core operations and supported by cash flow are sustainable. Earnings driven by one-time items, accounting adjustments, or unsustainable practices are not. Comparing adjustments across time reveals sustainability.
Summary
Comparing adjusted numbers across companies and time periods is essential for separating real earnings quality from accounting engineering. The method is straightforward: identify what each company adjusts for, apply uniform adjustment rules, normalize earnings, and compare. Companies with consistent normalized earnings supported by cash flow are credible; companies with growing adjustments, deteriorating GAAP results, and diverging cash flow are engaging in accounting manipulation. By doing this analytical work, investors see through creative accounting and identify companies whose adjusted metrics mask economic decline. This work is more time-consuming than trusting management's numbers, but it's the difference between sophisticated and naive investing.