Variant Perception

Figures converted from GBP at historical FX rates — see data/company.json.fx_rates for the rate table. Ratios, margins, multiples, and percentages are unitless and unchanged.

Where We Disagree With the Market

The single sharpest disagreement: the Mead Johnson exit that bull, bear and Stan are all waiting for is a write-down catalyst, not a re-rating catalyst. Consensus prices a binary — clean MJN sale at carrying value re-rates Core Reckitt to the 13.6x HPC peer median; no sale leaves the discount in place. The most-public sell-side MJN range (UBS: $5.4-8.9bn EV at 8-13x 2026 EBITDA, originally published in euros) sits 17-50% below Reckitt's $10.8bn carrying value, meaning the modal outcome is a $2.7-5.4bn impairment booked at signing alongside cash proceeds that fund one more special distribution but not the structural payout gap. Two related disagreements compound it: the HPC peer-median multiple bull and bear both anchor to is the wrong reference for a 15%-infant-nutrition mix, and the FY25 102% payout ratio is being framed as transient when the post-EH cash math says it is structural. Each disagreement is testable inside 18 months — and each cuts the same way.

Variant Perception Scorecard

The 68/100 variant strength reflects three converging disagreements that are each individually defensible from upstream evidence and that point in the same direction (lower fair value, slower catalyst, weaker carry). The 72/100 consensus clarity is high because sell-side rating distribution (Investing.com 10 buys / 0 sells; mean PT $88.36 versus a $64.27 print) and Stan's "Lean Long, Wait For Confirmation" verdict are both observable. The 74/100 evidence strength is anchored by primary disclosures (FCF conversion, payout ratio, intangible base, the Essential Home transaction structure) and by sell-side's own published MJN valuation range. The first resolution gate is the H1 2026 print at the end of July — six trading months away — but the central MJN-write-down disagreement only resolves on a disclosed transaction or a year-end CGU test, both of which sit in the H2 2026 / FY2026 reporting window.

Consensus Map

The consensus position is unusually crisp here because the sell-side ratings, Bull's own price-target build, and Stan's "Lean Long, Wait For Confirmation" verdict all converge on the same two pillars: the multiple discount is real-and-closeable, and the MJN exit is a positive-asymmetric catalyst within 12-18 months. The Q1 air-pocket framing is the lower-confidence consensus pillar — Kepler upgraded after the print, but the tape (5.6% sell-off, fresh death cross) and Morgan Stanley's December downgrade say not everyone bought it.

The Disagreement Ledger

Disagreement 1 — MJN as write-down, not re-rate. Consensus, articulated by both Bull and Stan, is that any defensible MJN transaction unlocks a 13-14x Haleon-tier multiple for Core Reckitt. The evidence we lean on says otherwise: UBS — currently the most explicit sell-side modeler of an MJN transaction — has bracketed enterprise value at €4.7-7.7bn (~$5.4-8.9bn) on 8-13x 2026 EBITDA, with Barclays' Warren Ackerman publicly flagging "valuation uncertainty and earnings per share accretion" as the question marks for any Danone bid. The $10.8bn carrying value sits above the high end of this published range. If we are right, a disclosed sale crystallises a $2.7-5.4bn impairment alongside cash proceeds that — net of debt assumed, deferred consideration, and any NEC indemnity escrow — most plausibly fund a single additional distribution but do not produce the bull's "second leg of the simplification trade." The market would have to concede that the $21.3bn intangible base was always the central balance-sheet risk and that adjusted EPS has been the wrong reference number through the cycle. The cleanest disconfirmation: a Danone (or PE infant-nutrition platform) cash bid above $9.4bn EV with no NEC indemnity escrow, accompanied by a CGU impairment of less than $670m at signing.

Disagreement 2 — Wrong peer set. Both Bull and Bear use the 13.6x HPC peer median as the gravity well. We think Reckitt is structurally HPC + 15% infant nutrition, and infant nutrition is an 8-13x category — UBS's own MJN bracketing confirms it, and Reckitt's MJN segment growth profile (FY25 LFL +3.8% vs Powerbrand EM +14.6%) confirms it operationally. A mix-weighted fair multiple is 11.8-12.6x, not 13.6x. If we are right, the bull's $87.53 price target — which assumes a peer-median rerate plus an MJN-exit credit — is structurally $9-13 too high; the consensus $79.24 Fintel mean target sits closer to the mix-adjusted ceiling. The market would have to concede that the "cheapest large-cap HPC" framing requires MJN to disappear before it becomes valid; until then, the discount is half what bull-and-bear-aligned tables show. The cleanest disconfirmation: MJN exits cleanly (Disagreement #1 resolves), at which point the peer-comp becomes the correct reference and the bull rerate math reasserts.

Disagreement 3 — Dividend is structurally uncovered. Consensus reads the FY24/FY25 102% payout ratio and the FCF conversion drop to 71% as transient — cash tax normalisation in FY26, EH stranded costs fade, restructuring tail rolls off, conversion returns to the 85-90% historical band. We think the math doesn't close at the smaller post-EH base. Essential Home contributed roughly $269-404m of FCF that is now gone; management's own normalised adj ETR is 24-25% (not the FY24 22.2% that included a tax tailwind that has reversed); capex stays at $808m+ for MJN regulatory upgrades and Wilson-NC plant readiness through FY27; the 5% progressive dividend at FY25's $2.85-per-share base is a $1.95bn annual call; even a halved buyback at $606m brings the total commitment to ~$2.56bn against a structural FCF base of $2.02-2.29bn. If we are right, the dividend either freezes in 2027 or the buyback stops, and at that point the income-mandate buyer base — which is the marginal holder at a 9.7x P/E and a 4.5% headline yield — sells. The market would have to concede that the carry-while-waiting thesis underwrites a payout that requires either (a) a clean MJN exit at carrying value (Disagreement #1) or (b) leverage migration above the 2.5x Moody's envelope. The cleanest disconfirmation: FY26 cash tax below $940m and FCF conversion above 85% on a like-for-like base, with the 5% dividend progression and the buyback both renewed at H1 results.

Evidence That Changes the Odds

The single most important row is row 1 — UBS's published MJN range is the rare case where a sell-side house has put numbers behind the qualitative "any defensible value" framing the rest of the bench uses. It is the strongest piece of evidence we have because it is the consensus's own modeling work, and it contradicts the consensus's own conclusion.

How This Gets Resolved

The resolution path runs in roughly this order: H1 2026 print (late July) tests Disagreements 2 and 3 directly; the 6 July NEC bellwether tests Disagreement 1 indirectly by setting the litigation discount in any MJN auction; an MJN disposition (or formal pause) in H2 2026 tests Disagreement 1 directly; the FY26 year-end goodwill test (March 2027) is the backstop that resolves Disagreement 1 even if no disposal lands. There is no "time will tell" here — every signal is observable in a filing, court docket, or capital-allocation announcement within 12 months.

What Would Make Us Wrong

The cleanest way the variant view breaks is a competitive MJN auction. UBS's $5.4-8.9bn range is a passive single-buyer model; if Danone, Nestlé and a private-equity infant-nutrition platform (Lone Star, Apollo and Advent were all shortlisted for Essential Home in February 2025) bid against each other, the high end of that range can be pierced. A clean all-cash sale at $9.4bn-plus with NEC indemnity capped at $670m would refute Disagreement 1 directly, validate the bull's "embedded option" framing, and trigger the second special distribution that backfills Disagreement 3. The probability is not zero — Danone is publicly engaged via Centerview Partners, has €9bn of liquidity per its latest URD, and has wanted Mead Johnson since 2014. The single best evidence we could see is a competitive process headline (two confirmed bidders) before mid-2026.

The second way we are wrong is on the peer-set call. If Mead Johnson exits before the H1 2026 print or before the next FY guide cycle, then the residual Core Reckitt is genuinely a 13-14x HPC peer-comp business and Disagreement 2 collapses immediately. We are betting MJN takes 12-18 months to clear and that the multi-quarter overlap of "still own MJN at 8-13x" with "Core deserves 13-14x" caps the blended fair value below the bull's target. A faster transaction sequence breaks our timing call, even if the price-discount thesis is right on the asset.

The third way we are wrong is on the dividend math. If FY26 cash tax surprises low ($875m or below), if working-capital normalises favourably as the EH carve-out fully unwinds, and if Fuel-for-Growth delivers another 150bps of fixed-cost takeout, FCF conversion can recover to 85% on a structural basis — at which point the $3.10bn payout commitment is closer to covered organically and Disagreement 3 weakens. Management has hit Fuel-for-Growth ahead of plan in FY24 and FY25; we should not assume they miss in FY26.

We are also conscious that the variant view is most useful inside a 12-18 month window. If we are wrong about MJN pricing, we will be wrong fast — a disclosed bid above $9.4bn would price into the stock the day it is announced. That asymmetry is worth respecting: the "watch for the wrong" signal is a single Reuters/Bloomberg headline, not a multi-quarter trend.

The first thing to watch is the 6 July NEC MDL bellwether verdict in N.D. Illinois — a defence sweep would compress the litigation discount embedded in any MJN auction and is the cleanest path the bull has to refuting Disagreement 1 before the H1 print arrives.