Forward P/E higher than trailing P/E on a company guiding earnings down means the trailing base is fake

Claim

The relationship between trailing and forward P/E encodes the market's earnings expectation:

• Earnings expected to grow → forward EPS higher → forward P/E < trailing P/E
• Earnings expected to fall → forward EPS lower → forward P/E > trailing P/E

So a forward P/E above trailing is not by itself an error — it is the normal signature of expected decline. The tripwire is the contradiction: forward above trailing on a company whose own guidance says earnings are flat-to-down, combined with a trailing multiple that looks anomalously cheap versus the name's history. That pattern means the trailing base contains something that is not repeatable, and the "cheap" trailing multiple is measuring it.

Evidence — NVO, 2026-08-05

Metric	Reported	Note
PE(ttm)	10.55x	Circulated as "cheapest multiple in a decade"
PE(fwd)	13.57x	Higher than trailing
FY2026 guidance	adjusted operating profit 0% to −6% CER	Guiding down

The inversion was visible in the fin.py snapshot with zero additional research. Chasing it found the cause immediately:

Novo Nordisk Q1 2026	DKK
Reported operating profit	59,618M — "+65% CER"
Adjusted operating profit	32,858M — "−6% CER"
Difference	26,760M
Cause	Reversal of a DKK 26.8B provision tied to the US 340B programme — non-cash, no cash-flow impact

That reversal sits inside TTM earnings. Reconstructed TTM diluted EPS was 26.19 DKK; stripping the after-tax reversal gives ~22 DKK. The true trailing multiple is ~12.5x, not 10.55x — and the corrected figure lands right next to the forward multiple, as it should.

Why this is dangerous rather than merely untidy: screens and watchlists rank on trailing P/E. A one-off gain therefore makes a name screen as its cheapest in the same period its underlying earnings are deteriorating — the exact inversion pitfall-dyt-inverts-when-price-caused-the-yield documents for dividend yield. Two different vendor metrics, same failure shape: the signal strengthens as the business weakens.

How to apply

1. Compare PE(fwd) to PE(ttm) on every snapshot. It is two numbers already on the page.
2. If forward > trailing, ask what guidance says. Guiding down → consistent, no flag, but the trailing multiple is still the wrong one to quote. Guiding up or flat while forward > trailing → the trailing base is inflated. Stop and find the one-off.
3. Where to look: reported vs adjusted operating profit in the company's own release. The gap is usually named explicitly — provision reversals, legal settlements, divestment gains, tax-valuation-allowance releases, unrealised equity marks.
4. Never quote a trailing multiple you have not reconciled to adjusted earnings on a name where the gap exceeds ~10% of operating profit.
5. The inverse case is equally useful: trailing P/E anomalously high with forward far below it can mean a one-off charge depressed the base — which makes a name screen expensive when it is not. Same test, opposite sign.

What would falsify this

A vendor computing forward EPS off a different fiscal year than assumed, which can invert the relationship without any one-off in the trailing base — see [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]]. Check that first: confirm which year the forward figure covers before concluding the trailing number is the broken one. On NVO the forward figure was approximately right and the trailing one was wrong; on TSM and QCOM it was the reverse.

Related

• [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] — the other half of this diagnostic; run it first.
• [[pitfall-dyt-inverts-when-price-caused-the-yield]] — same "signal strengthens as business weakens" shape.
• [[pitfall-tax-valuation-allowance-round-trip-breaks-eps]] · [[pitfall-unrealized-equity-marks-break-headline-pe]] · [[pitfall-yahoo-misclassifies-strategic-investment-gains]] — three other sources of the same inflated base.
• [[principle-primary-source-beats-vendor]] · [[pitfall-gate-cleared-by-adjective-not-arithmetic]]