Crane Company ($12B industrial) filed its 10-K on Feb 26. The stock trades at 27x forward earnings with 9 analysts covering it, 7 bullish. The nuclear thesis via Reuter-Stokes is consensus. This is a factor decomposition exercise, not an alpha opportunity.

What the 10-K actually says

Three new items the Q4 earnings call didn't surface:

Tariff exposure is explicit and specific. The risk factors section uses Trump-era language — "recently announced and potential additional tariffs" — applied to steel, copper, resin, and electronic components. The AAT segment "has been experiencing, and may continue to experience, supply chain disruptions from insufficient availability of certain components and raw materials." This isn't boilerplate. The 41% international sales mix creates reciprocal tariff exposure on both input costs and revenue.

PFT margin will decline in 2026. Both segments — AAT and PFT — expect "modest margin decline" from acquisition dilution. This matters because Reuter-Stokes (the nuclear sensing business everyone's excited about) sits in PFT, not AAT. The segment where nuclear growth is supposed to show up is the one facing margin compression. PFT backlog has been declining for three consecutive years ($379M to $376M to $360M), though to be fair, Reuter-Stokes only closed January 1, 2026 and isn't in the Dec 31 snapshot.

The dividend raise is real signal. 11% increase to $1.02/share — first raise since the Crane Holdings spinoff in April 2023. Management is signaling confidence in cash generation despite spending $1.15B on four acquisitions. Net leverage is 1.4x post-close. They're not stretching.

Factor decomposition

This is where CR falls apart as an investment for us:

Total idiosyncratic component: ≈25-30%. Below the 75% threshold by a wide margin. Even the nuclear thesis — the reason anyone gets excited about CR — is a sector factor shared across BWXT, CCJ, CW, FLS, CEG, and GEV. Our worldview has 11 evidence items across 7 tickers on nuclear buildout, average LR 1.94. CR is a diluted way to play that theme with loading 0.30 vs BWXT at 1.0.

Forward alpha

Raw return to analyst median target ($225): +10.1%. Subtract expected sector return (XLI running +35% 1Y, mean-revert to ≈10%): +0.1% excess. Apply edge percentage (20% generous — only acquisition execution): 0.02% annualized alpha.

Proportional sizing on a $2M portfolio: $210. Not a position. A rounding error.

Even with the most bullish analyst ($238): excess return 6.5%, edge-adjusted alpha 1.3%, position size $13.6K — sub-1% of portfolio. The math doesn't work.

Market-implied probability

The market is pricing ≈50% probability the full bull case delivers (nuclear + acquisitions = $9-10 EPS by 2027, stock at $225-250) against ≈40% muddle-through and ≈10% bear. Our 10-K findings nudge bear from 10% to 15% and bull from 50% to 45%. That moves fair value from $203 to $198 — a 3% disagreement with consensus on a $12B company covered by 9 analysts. Not actionable.

Insider activity

One real signal in the entire Form 4 trail: Director Jennifer Pollino bought 1,500 shares at $184.29 on January 29 — the morning after Q4 earnings — spending $276K of personal capital. Code P, open market purchase. She heard the results and backed them with her own money.

Everything else is comp mechanics. Alcala (incoming CEO) has never bought on the open market — every filing is RSU vests and tax withholding. His February 13 Rule 144 filing signals a planned future sale, directionally opposite to Pollino's buy. Mitchell's February 9 filing is his annual comp grant (PRSUs + options at $199.99 strike).

One director buying doesn't confirm the thesis. Management not buying alongside her doesn't kill it. Net insider signal: LR ≈1.1. Barely moves the needle.

The honest synthesis

CR is a well-run industrial compounder doing what it's supposed to do — growing revenue 8%, expanding margins, integrating acquisitions competently, and positioning for nuclear buildout. All of which is priced in at 27x forward earnings with 78% analyst bullish consensus.

The nuclear thesis is real but it's a sector factor, not company-specific alpha. Reuter-Stokes gives CR nuclear exposure, but it's the diluted version — loading 0.30 vs pure-plays at 0.8-1.0. If you want nuclear, buy the pure-play, not the conglomerate.

The 10-K added two bearish items (tariff headwinds, PFT margin compression) and one bullish item (backlog visibility + dividend raise). They roughly cancel out. No inflection. No mispricing.

We don't own it. We won't own it. The edge zone criteria ($3B cap, 10 analysts, 75% idio) are failed on every dimension. This is a well-covered consensus trade dressed in a nuclear suit.