# Most Stocks Destroy Wealth. Three Checks to Filter Them Out
> We ran the three income-statement checks analysts use to separate great companies from bad ones — tested against 100 years of data on 28,114 U.S. stocks.
- Author: Barebone Research, Barebone AI
- Published: 2026-04-15
- Canonical: https://barebone.ai/resources/how-to-read-an-income-statement-three-checks
- Publisher: Barebone AI (https://barebone.ai)
---
## The Uncomfortable Base Rate
In 2018, an Arizona State finance professor named Hendrik Bessembinder published a study in the Journal of Financial Economics that should be stapled to every new brokerage account. He measured the lifetime return of roughly 26,000 U.S. common stocks — every one listed since 1926 — and compared each against the most boring asset on Earth: the one-month Treasury bill, the short-term government IOU that is finance's stand-in for cash.
The results were brutal. Only **42.6%** of stocks beat T-bills over their lifetimes. The single most common lifetime outcome, rounded to the nearest 5%, was a **loss of 100%** — nearly 12% of every company ever listed went effectively to zero. The median stock survived about **seven and a half years** before being delisted, acquired, or buried.
His updated numbers, running 1926 through 2022, are blunter still. Of **28,114** stocks ever listed, **58.6% reduced shareholder wealth** over their time on the market. The market's entire **$55.1 trillion** in net wealth creation traces to the best-performing **4%** of companies — the other 96% collectively did no better than T-bills. Just **72 stocks** account for half of it.
Sit with that chart for a second. The stock market as a whole is a wealth machine. The average *stock* is a bad bet. Which means the first skill in investing is not picking winners — it is filtering out the losers fast enough that you never own them.
Professional analysts run that filter with one document and three checks. We used Barebone to pull the latest annual income statements of six of the largest companies in America and walk through exactly how the checks work — and, just as important, where they break.
## One Document, Three Checks
A great company does two things: it makes money today, and it makes more of it year after year. Everything else — the keynotes, the product demos, the CEO's podcast tour — is commentary.
The document that tells you whether those two things are true is the **income statement**: the official scorecard, filed with the SEC every year, of everything a company earned and everything it spent. Two lines matter most. **Revenue**, at the top, is all the money that came in. **Net income**, at the very bottom, is what survived after every cost — salaries, factories, marketing, interest, taxes. That position is why profit is called "the bottom line."
| Check | The question | Where to look | What passes |
|-------|--------------|---------------|-------------|
| 1 | Is it making money? | Net income, latest year | Positive |
| 2 | Is it making more over time? | Revenue *and* net income, last 3–5 years | Both rising |
| 3 | Is it keeping more of each dollar? | Net margin (net income ÷ revenue) | High for its industry, ideally rising |
Three checks, ten minutes, free data. Here is each one with the receipts.
## Check 1: Is It Actually Making Money?
Open the income statement. Look at the bottom line. Positive means profit. Negative means the company spends more than it brings in.
This sounds too basic to be a real filter. It is not. As of late 2023, roughly **40% of the companies in the Russell 2000** — the standard index of American small-cap stocks — had negative earnings. Four out of ten listed companies, not making money.
An unprofitable company is not automatically a bad investment. Every company starts out unprofitable, and some of Bessembinder's wealth-creating 4% spent years in the red before the flywheel caught. But an unprofitable company is a different *kind* of bet: you are betting on a story about future profits, not on profits. The modal outcome in the data — the 100% loss — is what happens when the story's "someday" never arrives.
Check 1 will not find you a single winner. It removes the single largest category of losers, which is worth more.
## Check 2: Is It Making More Every Year?
One year of profit can be luck. So pull the last three to five annual income statements — every public company posts them free on its investor-relations page — and look for two lines rising together: revenue growing *and* net income growing.
Why demand both? Revenue growth without profit growth means the company is buying its growth — every new customer costs more than they pay. Profit growth without revenue growth usually means cost-cutting, which works right up until there is nothing left to cut.
Now the correction to a myth you will hear constantly: that past growth is the best predictor of future growth. The academic record says otherwise. In a 2003 Journal of Finance study, Chan, Karceski and Lakonishok examined decades of operating performance across U.S. stocks and found **no persistence in long-term earnings growth beyond what pure chance would produce** — and that analysts' long-term growth forecasts were systematically too optimistic.
So why run Check 2 at all? Because a multi-year record certifies the *present*, not the future. A single great year can come from an asset sale, an accounting quirk, or a one-off demand spike. Five rising years in a row is much harder to fake — it proves the profit engine actually works. The track record verifies the machine. It does not promise what the machine produces next.
## Check 3: How Much of Each Dollar Does It Keep?
This is the check professional investors care about most, and it has a name: **margins**.
The arithmetic is one division. A company that brings in $100 of revenue and keeps $50 after all costs has a **50% net margin**. A company that brings in $200 but keeps only $20 has a **10% net margin**. The second company is twice the size. The first is five times the business.
The spread in the real world is enormous. Here are the six income statements we pulled, from the most recent fiscal year each company has filed:
| Company | Fiscal year | Revenue | Net income | Net margin |
|---------|------------|---------|------------|------------|
| NVIDIA | FY2026 (ended Jan 2026) | $215.9B | $120.1B | **55.6%** |
| Visa | FY2025 (ended Sep 2025) | $40.0B | $20.1B | **50.1%** |
| Microsoft | FY2025 (ended Jun 2025) | $281.7B | $101.8B | **36.2%** |
| Apple | FY2025 (ended Sep 2025) | $416.2B | $112.0B | **26.9%** |
| Amazon | FY2025 (ended Dec 2025) | $716.9B | $77.7B | **10.9%** |
| Walmart | FY2026 (ended Jan 2026) | $713.2B | $22.3B | **3.1%** |
Walmart and Amazon each push more than **$713 billion** through the top line — and keep about 3 cents and 11 cents of every dollar, respectively. Visa keeps **50 cents of every dollar** on a fraction of the revenue. A dollar of sales is not worth the same everywhere.
A fat margin is rarely an accident. It usually means customers pay a premium they could refuse but don't, and competitors can't undercut the price without losing money. And it compounds: every incremental dollar of revenue at a 50% margin drops five times more profit to the bottom line than the same dollar at 10%.
The evidence says the market has historically *underpaid* for this. Robert Novy-Marx's 2013 Journal of Financial Economics study found that profitability predicted the cross-section of stock returns roughly as well as the famous value effect — **profitable firms went on to outperform unprofitable ones even though they traded at richer valuations**. Quality costs more, and it has still been worth it.
## The Dream Version of Check 3
The best possible answer to Check 3 is not a high margin. It is a *rising* margin while revenue grows — proof that scale is making the business more profitable per dollar, not less. Analysts call this **operating leverage**: when revenue grows faster than costs, each new dollar of sales is more profitable than the last one.
The cleanest recent example is NVIDIA. Across five fiscal years, revenue grew from **$26.9 billion to $215.9 billion** — eight times — while the net margin climbed from **36.2% to 55.6%**:
Notice fiscal 2023, though. Revenue stalled and the margin collapsed from 36.2% to **16.2%** in a single year before the data-center boom arrived. Even elite businesses print ugly years. That is exactly why Check 2 demands several years of statements instead of one.
So, the full filter: making money? Making more of it every year? Keeping a bigger slice of each dollar as it grows? Three yeses is a good company. Anything else, keep looking — there are thousands of tickers and you only need a handful.
## Where the Three Checks Fail
Run honestly, the checks will also fail you in predictable ways. Four of them are worth knowing in advance.
**Net income can lie — in both directions.** In 2022, Amazon reported a net *loss* of **$2.7 billion** on **$514.0 billion** of revenue. Had the store stopped working? No — accounting rules forced Amazon to run a **$12.7 billion** unrealized markdown on its stake in Rivian, the EV maker, straight through the income statement. The underlying business was fine, and the bottom line snapped back immediately:
| Year | Amazon net income |
|------|------------------|
| 2021 | +$33.4B |
| 2022 | **-$2.7B** |
| 2023 | +$30.4B |
| 2024 | +$59.2B |
| 2025 | +$77.7B |
A mechanical Check 1 would have flunked one of the strongest businesses on Earth at precisely the wrong moment. When a bottom line spikes or craters in a single year, find the one-off item before you trust it.
**A thin margin can be the strategy.** Walmart's 3.1% looks feeble next to Visa's 50.1% — but Walmart *chooses* thin margins. Rock-bottom prices are the moat; the spread between industries is mostly physics, not quality. Comparing a retailer's margin to a payment network's tells you nothing. Compare a company against its own industry and its own history.
**The checks grade the company, not the stock.** Nothing on an income statement tells you whether the price you are being asked to pay is sane. A great company at an absurd valuation can still be a poor investment, and the market knows quality exists — it charges for it. Quality and price are two separate questions, and this filter only answers the first.
**Passing today is not passing forever.** The median listed company survived seven and a half years. Growth does not persist beyond chance. The three checks are a snapshot, which is why analysts re-run them every time a new annual report drops — and why a company that passed five straight years still gets re-examined on year six.
## What This Means When You Open the Filings
The three checks will not hand you the next member of Bessembinder's 4%. Nothing reliably does. What they do is structural: they keep you out of the 58.6% of stocks that destroyed wealth, the 40% of small caps earning nothing, and the modal outcome that rounds to a total loss.
A working version of the filter looks like this:
1. **Bottom line positive?** If not, you are buying a story — size the bet accordingly.
2. **Revenue and net income both rising across 3–5 years?** One great year is noise.
3. **Net margin high for its industry — and trending up as revenue grows?** Operating leverage is the dream; margin erosion while growing is the warning.
4. **Any single-year spike or crater?** Find the one-off item before believing it.
5. **Compared within the industry,** never across it.
Three yeses puts a company on the research list. It does not tell you what the stock is worth — a great business and a great price are different questions, and the second one is where most of the money is actually made or lost.
The income statement answers the first question in ten minutes. That is the trade most people never make: ten minutes of reading, against base rates that say the average stock quietly loses to cash.
---
*Data: Barebone | Sources: Bessembinder, "Do Stocks Outperform Treasury Bills?" (Journal of Financial Economics, 2018); Bessembinder, "Shareholder Wealth Enhancement, 1926 to 2022" (SSRN, 2023); Chan, Karceski & Lakonishok, "The Level and Persistence of Growth Rates" (Journal of Finance, 2003); Novy-Marx, "The Other Side of Value" (Journal of Financial Economics, 2013); NVIDIA, Visa, Microsoft, Apple, Amazon and Walmart annual filings, FY2025–FY2026 | Data as of April 15, 2026*