Free Cash Flow Lies Too: Three Adjustments Before You Trust It
Picture two companies that each report $1 billion of net income. Over the same year, one of them produces $1.3 billion of free cash flow. The other produces $400 million. Same headline profit, a $900 million gap in what an owner could actually take out of the business. Screens that sort stocks by earnings cannot see that gap, and it is the reason free cash flow is the first number I look up.
My view is simple to state. Free cash flow is the best single number for judging a business, because it is harder to manage than earnings, but the figure you copy off a data page is only a starting point. Three adjustments move it more than most people expect: stock-based compensation, the line between maintenance and growth spending, plus working capital. If you skip them, you are trusting an accountant’s second opinion as if it were the first.
Why cash is harder to flatter
Net income is built on accruals. Revenue is recorded when a sale is earned, not when the money arrives. Depreciation spreads the cost of a machine bought years ago across its useful life. Write-downs and one-off gains pass through the same statement. None of this is dishonest. It is what accrual accounting is designed to do, and it gives a smoother picture of each period. The side effect is that reported profit and real cash generation can drift apart for years without anyone breaking a rule.
Free cash flow starts one step closer to reality. You take cash from operations, which strips out the non-cash charges, and subtract capital expenditure, the money spent on property, plant and equipment. What is left is the cash the business produced that could pay down debt, fund dividends, buy back shares or finance an acquisition. That is the money an owner is ultimately paid from.
Because cash is harder to move around than an accrual, over long stretches a stock tends to follow it. I would not call that a law. Stocks can trade far from cash flow for years, in either direction. But when I see a company whose profit grows every year while free cash flow stays flat, I treat it as a question I need answered before I go further. I wrote about one case where free cash flow fell sharply while sales kept rising, in my look at Meta’s spending, and it is a good example of how the two numbers can tell different stories.
The first leak: stock-based compensation
Cash flow statements add stock-based compensation back to operating cash flow as a non-cash expense. Technically that is correct, because the company did not write a check. Economically it is a stretch. The company paid its employees, in shares instead of dollars. Those shares come out of the pockets of existing owners through dilution or through buybacks the company has to fund with real cash.
Here is why that matters. Suppose a software company reports $2 billion of free cash flow and $700 million of stock compensation. If the company then spends $700 million of that cash buying back shares only to keep the share count from rising, the cash never reached shareholders. It went to employees, with a detour through the market. Free cash flow after stock compensation is $1.3 billion, and that is the figure that describes what owners kept.
This is not a rounding error at some companies. For mature industrial and consumer businesses, stock compensation is a small fraction of free cash flow and I barely adjust for it. For technology and software names it can be a large share of the headline number, and I subtract it every time. When I read a buyback story, such as the one in my note on Apple’s buybacks and services margin, the first thing I ask is how much of the repurchase merely offsets issuance.
The second leak: what counts as capex
Capital expenditure is the line where management has the most discretion, and I read it with suspicion in both directions. A company can flatter free cash flow by underspending, letting plants and equipment age while the depreciation charge keeps falling behind reality. That works for a few years. Then a big replacement bill arrives.
The reverse also happens. A business investing heavily in new capacity looks weak on free cash flow today for reasons that have nothing to do with its profitability. If a company spends $3 billion on a plant that will earn a strong return for two decades, the free cash flow line understates its earning power. The honest question is how much of that spending is maintenance and how much is growth. Companies rarely split it cleanly. Some give an estimate in their annual report, and I take that estimate as a claim to test, not a fact.
There are also definition traps. Some firms buy equipment through leases, which shifts the spending out of the capex line and into financing. Others capitalize software development costs, so an expense that hits profit at a competitor becomes an investment outflow here. Two companies in the same industry can report free cash flow that is not comparable until you line up these choices. I check the notes for how each one treats leases and capitalized costs before I compare a pair.
| Adjustment | What the headline shows | What I do | When it matters most |
|---|---|---|---|
| Stock-based compensation | Added back as non-cash | Subtract it from free cash flow | Software and technology |
| Capex definition | One capital spending line | Split maintenance from growth, check leases | Capital-heavy and lease-heavy firms |
| Working capital | Swings with payment timing | Average several years, look for repeat inflows | Fast growers and distributors |
The third leak: working capital
Operating cash flow includes changes in working capital, meaning receivables, inventory and payables. A company that stretches the time it takes to pay suppliers holds on to cash longer, and that shows up as a cash inflow. Nothing about the business improved. It borrowed from its vendors.
A one-time boost like this can lift free cash flow by hundreds of millions in a single year, and it reverses when the payment schedule normalizes. Distributors, retailers and rapidly growing companies are the usual suspects. A retailer that adds stores can show a big payables inflow because its suppliers are shipping ever more product on credit terms. It is real cash, but it depends on the growth continuing. The moment expansion slows, the inflow shrinks or turns negative.
I watch for a specific pattern. If operating cash flow is much higher than net income in a given year, and the difference comes mostly from a payables or deferred revenue line, I do not annualize it. One year of that pattern is timing. Four years in a row means the business has a structural financing advantage, and I count that as a real strength, provided the customers or suppliers do not change the terms.
The test I would run on any name
The fix for all three leaks is dull. It also works.
Pull four years of cash flow statements. Average the free cash flow, which smooths working capital and lumpy capex. Subtract average stock-based compensation. Then set the result beside average reported net income.
If the two are within about 20 percent of each other, I read the profits as real. If cash after stock compensation runs at less than 70 percent of net income across four years, I stop and look for the reason. Those thresholds are my own rules of thumb, not a standard. I would loosen them for a company in the middle of a big build-out.
A concrete version makes the point. Say a company reports net income of $500, $560, $610 and $680 million across four years, an average of about $588 million. Its free cash flow was $410, $450, $380 and $520 million, an average of $440 million, and stock compensation averaged $120 million. Cash after stock compensation is $320 million, or about 54 percent of profit. Reported earnings have climbed 36 percent while owner cash has not shown a trend, and I would want to know why before paying a premium multiple. Those numbers are made up to show the arithmetic, so do not look for the company.
Where free cash flow fails on its own
I should be fair to earnings here. Free cash flow is a poor guide for banks and insurers, where cash flow is dominated by deposits, loans and claims that behave nothing like a factory’s inventory. For those I use other tools, like the side-by-side approach in my table of four big US banks. It is also a weak guide for young companies that are spending ahead of revenue on purpose. Negative free cash flow can be a good sign there, if the spending earns returns you can see.
Cyclical businesses are the other trap. A steelmaker or a chip equipment maker can show huge free cash flow at the top of a cycle because inventories are shrinking and prices are high, and a weak number at the bottom when the opposite is true. One year of free cash flow says little. That is why I average.
And there is a counter-case to my own advice, which I take seriously. If a company earns a high return on every dollar it reinvests, then heavy spending that depresses free cash flow is exactly what I want it to do. A rule that punishes it would keep me out of the best compounders. I try to solve that by asking what return the last few billion of capex earned, not just how big it was. When that return is falling, I distrust the growth story. When it is stable, low free cash flow is fine.
What I would look at first
Free cash flow is a better anchor than earnings. It is not a guarantee, and the risk is trusting it blindly. The habit I would build is a three-line check every time: cash from operations minus capex, minus stock compensation, then a glance at whether working capital is doing the heavy lifting. Once that becomes routine, most misleading figures show up within a few minutes, before they turn into a position. It is the same discipline behind the price-to-cash arguments in my look at what Microsoft’s multiple asks.
Take one company you already follow, pull the last four annual cash flow statements and run the adjustment. If cash after stock compensation stays under 70 percent of reported profit, that is your cue to read the footnotes.
Analysis and opinion only, not investment advice. Figures to check come from the cash flow statements that US-listed companies file with the SEC, searchable on SEC EDGAR, and the EDGAR overview explains what each filing type contains. The company numbers in my examples are hypothetical. The thresholds are my own rules of thumb, reviewed on September 22, 2026.