Check the deferred-tax line before believing any free-cash-flow surge

Claim

Before annualising, extrapolating, or scoring on an FCF-margin jump, read the deferred income taxes line in the cash-flow statement. A tax-timing swing lands inside operating cash flow and is indistinguishable from operating improvement at the FCF level.

What happened

analysis_notes.md §1 makes FCF the backbone — "FCF is harder to fake than earnings." That is true about fraud and false about timing. Intuit FY2026:

	Headline	Underlying
FCF	$8.617B, +41.7%	~+13%
Driver	—	~65% a deferred-tax swing (OBBBA §174 R&D capitalisation reversal)

The +41.7% was quoted forward across two reports and a watchlist entry as evidence the business had inflected. It had improved — 13% is a good year — but not by 42%, and the difference is exactly the gap between a re-rating and a continuation.

The same figure produced a health score of 8.5 where owner earnings support roughly 7.5.

Why FCF is the wrong place to catch it

Deferred taxes are a non-cash add-back inside operating cash flow. OCF − capex therefore carries the whole swing, and every downstream metric inherits it: FCF margin, FCF/share, FCF yield, FCF payout, EV/FCF. Nothing in the FCF number itself signals it.

The cross-checks that do catch it:

Compare FCF growth to net-income growth. A large divergence in either direction is the flag. (The inverse case — FCF growing far slower than net income — is the KLAC signature: FY26 FCF +0.5% against net income +18.9% on a working-capital build.)
Read deferred income taxes as a percentage of the FCF delta. Above ~25%, quote the clean figure alongside the headline.
Ask whether a tax-law change landed in the period. OBBBA §174 affected the whole US software sector in the same fiscal window, so this is not a single-company quirk — expect correlated false positives across a sleeve.

The rule

A single year's FCF growth is not evidence of anything until the deferred-tax line is read. analysis_notes.md §0 already says "trend over snapshot"; this is the specific mechanism by which a one-year FCF snapshot lies.
When reporting a headline FCF jump, report the clean figure next to it. "FCF +41.7% headline, ~+13% clean" costs four words and prevents the error.
Applies symmetrically to a collapse: check whether the fall is a timing reversal before writing down a thesis.

Same family as [[pitfall-tax-valuation-allowance-round-trip-breaks-eps]] (Dynatrace — a deferred- tax valuation-allowance release destroying the reported EPS series) and [[pitfall-single-quarter-fcf-read-as-ttm]] (Intuit again, from the opposite direction: a single quarter read as a trailing year made FCF look like it had fallen). Intuit has now produced two opposite FCF misreadings in three months, which is itself the signal: this issuer's cash-flow statement needs the line-by-line read, not the summary.

What would falsify this

FY2027 FCF growing at or above the headline FY2026 rate without a tax tailwind, which would show the deferred-tax component was a pull-forward of real cash rather than a distortion of the growth rate. The Q1 FY27 print (~2026-11-20) is the first read.

History