SYNTHOS RESEARCH

Marsh & McLennan Companies MRSH

Financial Services · Insurance - Brokers · Synthos Deep Dive · 2026-08-04

$192.64
Watch
—
—

The Overview

When a large company buys insurance, it does not phone an insurer. It hires a broker, who knows which insurers will write which risks at what price, structures the programme, and negotiates on the client's behalf. Marsh is the largest such broker in the world. It also owns Guy Carpenter, which does the same job for insurers buying reinsurance, and Mercer, which advises companies on pensions, health benefits and pay.

It is an unusually good business. The revenue recurs — corporate insurance renews annually and nobody switches brokers casually. It requires almost no capital: Marsh spent $291 million on equipment last year against $5.3 billion of cash from operations. And it earns about 26% a year on shareholders' money.

So why has the stock done nothing for a year while the market rose 24%?

Three reasons, and they are all identifiable.

First, a lawsuit. In 2014 a Marsh subsidiary in London arranged trade-credit insurance for a firm called Greensill Capital, which collapsed spectacularly in 2021. Marsh has been in litigation about it ever since. In the first quarter of this year it set aside $425 million for the estimated cost. That single charge turned a year of rising profits into a year of falling reported profits.

Second, prices. Brokers earn commissions that move with insurance premiums, and insurance premiums have stopped rising. The insurance side of Marsh grew only 3% last quarter after stripping out currency and acquisitions. Its consulting side grew 8%.

Third, indigestion. Marsh spent about $8.5 billion in 2024 buying McGriff, an American brokerage, and it is still absorbing it — $89 million of integration costs in the first half alone.

Meanwhile the company is returning a lot of cash: it bought back $1.5 billion of its own stock in six months and just raised the dividend 10%. Together that is about 5.4% of the company returned each year.

At $191 you pay about 17 times what analysts expect the company to earn in 2027. For a business of this quality that is cheap by its own history. We think it is worth about $208, or 9% more.

That is not enough for us to call it a buy. It is enough to put it on a list and set a price.

Putting a number on it: our fair-value estimate is $208 against a current price of $192.64 — real upside if our numbers are right.

Our summary metrics

Downside Risk (lower = safer)5/10Moderate

"Rated 5. The business risk is low and the balance-sheet and litigation risks are real. Supports: revenue is overwhelmingly recurring commission and fee income across 130 countries with roughly 95,000 colleagues; beta is 0.603, the second-lowest in this batch; interest coverage is 6.28x; the effective tax rate is a stable 24.2% for the quarter and 24.6% for the half; and the company signed a new $4.25 billion five-year revolving credit facility on 2026-06-02, expiring June 2031, upsized from the $3.5 billion facility it replaced. Against that: leverage is meaningful and the vendor overstates it. Filed long-term debt was $18,891M at 2026-06-30 and $18,320M at 2025-12-31 with a current portion of $646M and $1,267M respectively, so filed total debt at year-end 2025 was $19,587M against $2,687M of cash — net debt of approximately $16,900M, or 2.29x FY2025 EBITDA of $7,395M. Goodwill of $24,337M and intangibles of $5,212M are 48.4% of total assets against $15,100M of equity, so tangible book value per share is MINUS $27.85 — a rolled-up acquirer with no asset backing whatsoever. The Greensill litigation is the specific, dated, quantified risk: Marsh Ltd. placed trade-credit insurance for Greensill Capital from 2014; a $425 million estimated liability and legal expense was recorded in the first quarter of 2026; a settlement with Greensill Bank and its insolvency administrator was reached in June 2026; and claims by Credit Suisse in the Australian proceedings remain outstanding with the recorded liability reflecting only management's best estimate. Cyclical risk is the softening property-and-casualty rate environment, which the company names first in its own forward-looking-statement list ('changes in insurance premium rates') and which shows in underlying Risk and Insurance Services growth of 3% against Consulting's 8%."

Growth Quality5/10Moderate

"Rated 5 — steady, decelerating, and better than the GAAP line suggests. Revenue: $17,224M (2020), $19,820M (2021), $20,720M (2022), $22,736M (2023), $24,458M (2024), $26,981M (2025) — a 9.4% compound rate over five years, though the last two years include the McGriff acquisition, which consumed $8,453M of cash in FY2024. The June 2026 quarter grew revenue 6% reported and 5% underlying; the six months grew 7% reported and only 4% underlying, which is the honest organic rate. The segment split matters: Risk and Insurance Services grew 4% reported and 3% underlying in the quarter, while Consulting grew 10% reported and 8% underlying, with Marsh Management Consulting alone up 15% reported and 13% underlying. The company's largest business is growing at its slowest rate in years and its smallest is accelerating. Adjusted earnings per share grew 9% in the quarter to $2.96 and 8% in the half to $6.25; GAAP earnings per share fell 5% in the half to $4.99 because of the Greensill charge. Consensus wants adjusted EPS of $10.448 in FY2026 (14 analysts), $11.401 in FY2027 (14) and $12.481 in FY2028 (4) — 7.2%, 9.1% and 9.5% growth against a FY2025 actual we reconstruct at $9.75. The Thrive programme adds a mechanical lever: approximately $500 million of cost over three years for approximately $400 million of annualised savings, evenly distributed, launched in the third quarter of 2025. A 5, not higher, because 3% underlying growth in the core is what it is, and not lower, because the earnings line is compounding at 8-9% with a shrinking share count."

Exponential Potential2/10Low

"Rated 2. This is a people business. Approximately 95,000 colleagues advise clients in 130 countries, and revenue scales with headcount, acquisitions and insurance premium rates rather than with any curve. The company itself names the one genuinely non-linear force in its risk language, and it names it as a threat rather than an opportunity: 'our ability to compete effectively and adapt to competitive pressures and market changes in each of our businesses, including from disintermediation as well as technological change, digital disruption and other types of innovation such as artificial intelligence.' Against that, the Business Client Services unit formed alongside the Thrive programme is explicitly an attempt to run the other way — it 'brings together operations and technology teams across the Company' to 'centralize investments in operational excellence, data, artificial intelligence and other analytics,' targeting roughly $400 million of annualised savings on a cost base near $21 billion. That is a 1.9% efficiency programme, not a re-platforming. The one structurally interesting asset is the placement franchise itself: Marsh sits between clients and the entire property-and-casualty capital market, and that position is genuinely hard to replicate. It is also, by construction, an intermediary position, and intermediaries are what disintermediation removes. A 2: a superb, durable, linear business with one identified non-linear risk pointed at it."

Fair value$208 $158–$248
What does “fair value” mean?

Fair value is Synthos’s estimate of what one share is worth today, based on our model of the company’s future cash generation and the risks to it. It is not a price target or a prediction of where the stock trades next quarter — it’s the price where we think risk and reward are balanced. Above it, you’re paying for outcomes better than our base case; below it, the market is offering a margin of safety.

The model’s inputs and weightings are proprietary — they’re the product. What we publish is the output, plus an independent, public-math cross-check further down the page (“Check our number”) so you can judge it for yourself.

What do the 5 tiers mean? (Core · Tactical · Watch · Hold · Avoid)
Buy — CoreOwn it as a foundation — start or add now, size it for years, let dips be gifts.
Buy — TacticalGood price + confirmed trend + a defined exit — buy the setup, not a marriage.
WatchWe want the business, just not at this price/setup — act only when the listed trigger hits.
HoldFine to keep if you own it — no reason to buy more; new money does better elsewhere.
AvoidDon't own it — the problem is the business or the expectations, so a cheaper price won't fix it.

The Road Ahead

What we expect to matter in each window, and the evidence that would prove us wrong.

Short term 0-6 months

Neutral
Driver
"A sharp recovery already in progress, and an entry that is no longer cheap. Marsh closed 2026-08-04 at $191.45, DOWN 0.72% on the day, but 10.8% above a 50-day moving average of $172.71 and 8.1% above a 200-day average of $177.14 — an extended position relative to both. RSI is 66.5 and MACD is +5.51, the second-strongest momentum reading in this batch. Three-month return is +14.0% against SPY's +7.6%. But the twelve-month return is MINUS 3.2% against SPY's +24.3% — a 27-point deficit and the worst relative twelve months in this batch — with a maximum drawdown from peak of 21.6%. So the shape is: a bad year, then a violent seven-week recovery that has taken the stock from below its 200-day average to 8% above it. The June-quarter print on 2026-07-21 beat adjusted EPS by 2.8% and is the proximate cause. What tempers the stance in both directions: the stock is 9.4% below its 52-week high of $211.21 and 21.7% above its low of $157.32, so it sits mid-range after a round trip, and the next print is 72 days away."
What we’re watching
"The 2026-10-15 print against consensus adjusted EPS of $1.97 and revenue of $6,664M. Within it, three numbers rather than the headline: UNDERLYING revenue growth, which was 5% in the June quarter but only 4% for the six months and only 3% in Risk and Insurance Services — the single series that determines whether this is a cyclical trough or a structural slowdown; the adjusted operating margin, which was 29.3% against 29.5% a year earlier and 30.5% for the half against 30.7%, i.e. flat to slightly DOWN despite the Thrive savings; and any further Greensill development, because the $425 million recorded in the first quarter reflects the settlement with Greensill Bank plus management's best estimate of remaining Credit Suisse claims in the Australian proceedings, and estimates of litigation are not settlements. Also watch the buyback pace against $1.5 billion spent in the first half, and whether the Thrive cost line of $103 million for the half stays on the roughly $500 million three-year path."
Confidence
Medium

Medium term 6-24 months

Tailwind
Driver
"The medium-term case is a re-rating case and it rests on arithmetic rather than on hope. At $191.45 the stock is 18.3x the FY2026 consensus of $10.448 and 16.8x FY2027's $11.401. For a business earning 26.5% on equity with 30.5% adjusted operating margins, recurring fee revenue, a 0.60 beta and a twenty-year record of growth, that is a low multiple by its own history, and the de-rating is traceable to three identifiable and largely finite causes: the Greensill charge, the McGriff integration, and a decelerating property-and-casualty rate environment. Two of the three resolve. Meanwhile the capital return is doing real work: $1.5 billion of stock repurchased in the first six months of 2026 (8.7 million shares, against 478 million outstanding at 2026-06-30, down from 492 million a year earlier), and a dividend raised 10% on 2026-07-08 to $0.990 per quarter, $3.96 annualised, a 2.07% forward yield. Together that is roughly 5.4% of market capitalisation returned annually. Add consensus adjusted EPS growth of 7-9% and the expected return is 12-14% before any multiple move. Thrive contributes approximately $400 million of annualised savings against roughly $21 billion of costs by 2028."
What we’re watching
"Whether underlying growth reaccelerates. Risk and Insurance Services at 3% underlying is the problem: it is roughly two-thirds of revenue and it is growing at half the rate of Consulting. Property-and-casualty rate softening is the likely cause and it is cyclical, but the alternative explanation — that scale brokers are losing share to specialists, or that pricing transparency is compressing commissions — is structural and would not resolve. Watch: underlying growth in Marsh Risk specifically (4% in the quarter, with US/Canada at 4% and International at 5%); the adjusted operating margin, which must expand for the Thrive savings to be visible; whether the Greensill liability proves sufficient; the pace of bolt-on acquisition (5 deals for $181 million in 2026, a fraction of the $8,453 million spent on McGriff in 2024); and net leverage against the new $4.25 billion revolver's quarterly-tested coverage and leverage covenants."
Confidence
Medium

Long term 2+ years

Neutral
Driver
"Long-run, Marsh occupies one of the better structural positions in financial services and faces one of the more credible long-run threats to it. The position: an intermediary between corporate risk and the entire property-and-casualty capital market, with the scale, data and relationships to place complex programmes that no single insurer will write alone — a business that has compounded for a century and whose customers are large, sticky and regulated into buying. The threat is the one the company itself lists: disintermediation, digital disruption and artificial intelligence. A broker earns its fee for matching, structuring and advising, and each of those three is a task that improves with better information processing. That does not mean the fee disappears; the P&I club example in the knowledge base's discarded lane is a reminder that buyers organise collectively when intermediation gets expensive. It does mean the terminal multiple should carry a discount that a pure toll road would not. The rebrand to a single Marsh identity, the Thrive programme and the formation of Business Client Services are all the same bet: that consolidating the operating model and centralising data and analytics investment gets ahead of it."
What we’re watching
"Whether the fee pool holds. Concretely: commission and fee revenue per client relative to premium placed, which the company does not disclose and which would be the first place erosion appears; whether Consulting — growing 8% underlying against Risk and Insurance Services' 3% — becomes a larger share of the mix and whether that changes the multiple the market pays; whether Thrive delivers the promised $400 million of annualised savings by 2028 or becomes a rolling programme like several in this batch; whether the rebrand to a single Marsh identity is a cosmetic exercise or is followed by an actual reorganisation of the four historic operating companies; and whether the acquisitive model resumes at McGriff scale, because a company with negative tangible book value and $24.3 billion of goodwill has already spent its balance-sheet flexibility once."
Confidence
Low

Exponential Potential

Exponential Potential2/10Low

"Rated 2. This is a people business. Approximately 95,000 colleagues advise clients in 130 countries, and revenue scales with headcount, acquisitions and insurance premium rates rather than with any curve. The company itself names the one genuinely non-linear force in its risk language, and it names it as a threat rather than an opportunity: 'our ability to compete effectively and adapt to competitive pressures and market changes in each of our businesses, including from disintermediation as well as technological change, digital disruption and other types of innovation such as artificial intelligence.' Against that, the Business Client Services unit formed alongside the Thrive programme is explicitly an attempt to run the other way — it 'brings together operations and technology teams across the Company' to 'centralize investments in operational excellence, data, artificial intelligence and other analytics,' targeting roughly $400 million of annualised savings on a cost base near $21 billion. That is a 1.9% efficiency programme, not a re-platforming. The one structurally interesting asset is the placement franchise itself: Marsh sits between clients and the entire property-and-casualty capital market, and that position is genuinely hard to replicate. It is also, by construction, an intermediary position, and intermediaries are what disintermediation removes. A 2: a superb, durable, linear business with one identified non-linear risk pointed at it."

What could take this further than the base case — and where the market may be underpricing it. Full forward-growth and acceleration math in Deeper analysis, §4 below.

Deeper analysis

Technicals, fundamentals, valuation, and the full expert-claim evidence panel — the detail behind the numbers above.

Check our number Market-implied growth ≈ 17%/yr This isn’t how we calculate fair value — it’s public math you can verify yourself. To justify today’s $193, earnings would have to compound roughly 17% a year for 10 years (9% discount rate). Analysts forecast ~11%/yr, so the market is pricing in MORE than what the Street expects.

Reference table

Street consensus$203.71 (+6.4%) · median $201 · high $234 (+22.2%) · low $180 (−6.0%) · 1 strong buy / 10 buy / 22 hold / 1 sell across 34 analysts · vendor consensus label Hold
Valuation18.3x FY2026E adjusted ($10.448) · 16.8x FY2027E ($11.401) · 15.3x FY2028E ($12.481) · 23.2x trailing GAAP · EV/EBITDA 19.5x · P/B 6.08x · price/tangible book: NOT MEANINGFUL — tangible book is MINUS $27.85 per share
June quarter — filing-verifiedRevenue $7,404M, +6% (+5% underlying) · GAAP operating income $1,899M, +4% · adjusted operating income $2,166M, +5%, margin 29.3% (vs 29.5%) · GAAP EPS $2.63 · adjusted EPS $2.96, +9%
Six months — and this is why the stock de-ratedRevenue $15.0bn, +7% (+4% underlying) · GAAP operating income $3,653M, DOWN 5% · adjusted operating income $4,579M, +7% · GAAP EPS $4.99 vs $5.23 · adjusted EPS $6.25, +8% · the bridge is a $425 million Greensill litigation charge taken in Q1 2026
Capital return$1.5 billion of stock repurchased in six months (8.7M shares; 4.5M for $750M in Q2 alone) · dividend raised 10% on 2026-07-08 to $0.990/quarter, $3.96 annualised, a 2.07% forward yield · combined shareholder yield approximately 5.4%
Data findings — both filing-verifiedtotalDebt of $21,449M is 9.5% ABOVE the filed $19,587M, because $1,862M of lease obligations are added to debt the balance sheet already reports separately. seg_prod FY2025 reports one line at 116.7% of TOTAL COMPANY REVENUE and sums to 162.8% — the worst segment failure recorded in this programme
Technicals+10.8% above the 50-DMA ($172.71), +8.1% above the 200-DMA ($177.14); −9.4% from the 52-week high of $211.21, +21.7% above the low of $157.32; RSI 66.5; MACD +5.51; 12-month −3.2% vs SPY +24.3%; max drawdown −21.6%
Knowledge baseEMPTY. Zero name-level claims. 13 raw hits, 4 case-sensitive survivors, none carrying the company as an entity. The entire insurance-brokerage industry is absent from this store

What the experts actually said

No independent expert claims in the Synthos knowledge base yet for MRSH — this dive is fundamentals- and technicals-driven, not panel-driven.

Price & moving averages 12 months · 50 & 200-day averages · 52-week range

152171189208227Aug '25Nov '25Jan '26Apr '26Jun '26Aug '2652w hi $207Price 19350-DMA 182200-DMA 17852w lo $157

Solid line = price · dashed line = 50-day average · dotted line = 200-day average · the two thin horizontal lines mark the 52-week high and low. Price above both averages is an uptrend.

Data summary: last close $192.64, 6% above the 50-day average ($182), 8% above the 200-day average ($178) — an uptrend. 7% below the 52-week high of $207, 22% above the 52-week low of $157.

Bollinger Bands 20-day average ± 2 standard deviations

152169185202218Aug '25Nov '25Jan '26Apr '26Jun '26Aug '26Price 19320-day avg 191

The shaded band widens when the stock gets more volatile. Riding the upper edge = strong momentum (sometimes stretched); the lower edge = weak / potentially oversold.

Data summary: price $192.64 is currently inside the band (band $186–$196).

RSI (14) momentum gauge · 0–100

705030Aug '25Nov '25Jan '26Apr '26Jun '26Aug '26RSI 57.6

Above 70 (overbought zone, shaded) = overbought, below 30 (oversold zone, shaded) = oversold. Currently 58.

MACD 12 / 26 / 9 · trend & momentum

0Aug '25Nov '25Jan '26Apr '26Jun '26Aug '26signal 2.8MACD 2.5

The MACD line crossing above the signal line (bars flip to the up color) = momentum turning up; crossing below (bars flip to the down color) = turning down. Bar height = the size of that gap.

Data summary: MACD is currently below its signal line by 0.31, negative momentum.

Relative performance vs S&P 500 & its sector (XLF (sector)), set to 100 a year ago

738598111123Aug '25Nov '25Jan '26Apr '26Jun '26Aug '26S&P 500 119XLF (sector) 108MRSH 93

Solid = MRSH · dashed = S&P 500 · dotted = XLF (sector). A rising line means it is beating that benchmark — the sector line shows whether it is a leader or laggard within its own group.

Forward revenue & earnings actual → estimate · "FY" = fiscal year, "E" = estimate

09182736$20BFY21EPS $6$21BFY22EPS $7$23BFY23EPS $8$24BFY24EPS $9$27BFY25EPS $10$29BFY26EEPS $10$30BFY27EEPS $11$32BFY28EEPS $12

Darker bars = actual results, brighter = analyst estimates. Taller bars to the right = expected growth.

Key stats an RIA wants

Price$192.64
Market cap$92B
P/E trailing23×
P/E FY26E / FY27E18× / 17×
EV / Sales4.0×
EV / EBITDA19.6×
Gross margin42.4%
Net margin14.2%
Dividend yield1.92%
Beta0.579
52-wk range$157 – $207
RSI(14)54
50 / 200-DMA$182 / $178
12-mo return+-7% (SPY +19%)
Street target$203 ($182–$234)
Analyst grades10 Buy · 22 Hold · 1 Sell
FMP ratingB+
Next earnings2026-10-15 (Q3 2026 earnings, 72 days away; vendor consensus adjusted EPS $1.97 and revenue $6,664M, implying +4.9% revenue growth on the year-ago $6,351M). Second-quarter results were released 2026-07-21, fourteen days before this dive, and beat consensus adjusted EPS by 2.8% ($2.96 against $2.88) on revenue 1.7% above estimate.

1. What the business is, and what it now calls itself

The company has rebranded. The June-quarter press release opens: "Marsh (NYSE: MRSH), a global leader in risk, reinsurance and capital, people and investments, and management consulting, today reported financial results..." and the About section reads "Marsh (NYSE: MRSH) is a global leader in risk, reinsurance and capital, people and investments, and management consulting, advising clients in 130 countries. With annual revenue of $27 billion and more than 95,000 colleagues..."

The SEC registrant remains "Marsh & McLennan Companies, Inc." — that is the name on the 10-Q, the 8-K signature blocks and the reconciliation schedules — but the market-facing identity, the ticker and the operating-unit names have all changed. The historic Marsh / Guy Carpenter / Mercer / Oliver Wyman four-company structure now appears in the release as Marsh Risk and Marsh Management Consulting, and the restructuring programme that produced the change is described as focusing "on our brand strategy." The 8-K cover pages also list NYSE Texas alongside the New York Stock Exchange.

Two reportable segments, June 2026 quarter (all filing-verified from the 8-K of 2026-07-21):

SegmentQ2 2026 revenuereportedunderlyingGAAP op. incomeAdjusted op. incomeAdjusted marginprior year
Risk & Insurance Services$4.8bn+4%+3%$1,478M (+2%)$1,700M (+3%)35.3%35.6%
of which Marsh Risk$4.1bn+6%+4%————
Consulting$2.6bn+10%+8%$502M (+10%)$533M (+11%)20.5%20.2%
of which Marsh Mgmt Consulting$1.0bn+15%+13%————
Corporate/Eliminations———($81M)($67M)N/A—
Total$7,404M+6%+5%$1,899M (+4%)$2,166M (+5%)29.3%29.5%

Two observations that carry the whole medium-term question.

First, the core is growing at 3% and the smaller segment at 8%. Risk and Insurance Services is roughly two-thirds of revenue and its underlying growth was 3% in the quarter (and 3% for the six months). Consulting grew 8% underlying, with Marsh Management Consulting up 13%. Within Marsh Risk, US/Canada underlying growth was 4% and International 5% — EMEA 5%, Asia Pacific 5%, Latin America 8% — so the American core is the slowest part of the slowest segment. The obvious explanation is a softening property-and-casualty rate cycle, which is cyclical; the alternative is commission compression, which is not. The filings do not settle it and neither do we.

Second, adjusted margins are FLAT, not expanding. Consolidated adjusted operating margin was 29.3% against 29.5%; for the six months, 30.5% against 30.7%; Risk and Insurance Services 35.3% against 35.6%. Consulting is the only line where the adjusted margin improved (20.5% from 20.2%). The Thrive programme is targeting roughly $400 million of annualised savings and has cost $103 million in the first half — and none of it is yet visible in the adjusted margin.

2. The Greensill charge — the $425 million that explains the share price

From the 10-Q filed 2026-07-21, verbatim:

> "From 2014, Marsh Ltd. was engaged by Greensill Capital (UK) Limited and its affiliates as its insurance broker. Marsh Ltd. placed a number of trade credit insurance policies for Greensill. On March 1, 2021, Greensill filed an action against certain of its trade credit insurers in Australia seeking a mandatory injunction compelling these insurers to renew coverage under expiring policies. Later that day, the Australian court denied Greensill's application. Since then, a number of Greensill entities have filed for, or been subject to, insolvency proceedings, and several litigations and investigations have been commenced in the U.K., Australia, Germany, Switzerland and the U.S."

> "In November 2024 and March 2025, Greensill Bank AG (in insolvency)... added Marsh Pty Ltd. and Marsh Ltd., respectively, as parties to the Australian proceedings. In June 2026, Marsh Ltd., Marsh Pty Ltd. and other parties reached a settlement with Greensill Bank and its insolvency administrator."

> "In the first quarter of 2026, the Company recorded an estimated liability and legal expenses of $425 million related to the Greensill litigation... The recorded liability reflects the settlement with Greensill Bank and management's best estimate of the costs and losses associated with the remaining claims in the Australian proceedings brought by Credit Suisse."

Three things must be said about this and they run in different directions.

It is large and it is the reason the reported numbers look bad. $425 million against six-month GAAP operating income of $3,653 million is 11.6%. Without it, six-month GAAP operating income would have grown roughly 6% instead of falling 5%, and six-month GAAP EPS would have been approximately $5.65 instead of $4.99 (our arithmetic: $425M pre-tax at the disclosed 24.6% six-month effective tax rate is approximately $320M after tax, or $0.66 per share on 484 million diluted shares). That is the entire difference between "earnings fell" and "earnings rose 8%."

It is a settlement plus an estimate, not a settlement. The Greensill Bank matter is settled. The Credit Suisse claims in the Australian proceedings are not, and the recorded liability is explicitly "management's best estimate." A litigation estimate is not a resolved liability and this dive does not treat it as one. The kill criteria in Section 9 include a further Greensill provision for that reason.

It also removes an overhang that has been running for five years. A matter that has been live since March 2021, across five jurisdictions, has now been quantified for the first time and partly settled. For a market that has de-rated this stock by roughly a quarter, quantification is worth something even when the number is large.

3. The quarterly record — read the adjusted line, and know what is in it

QuarterRevenueYoYGAAP op. incomeGAAP diluted EPSAdjusted EPSvs est
Q2 2024$6,221M—$1,642M$2.27——
Q3 2024$5,697M—$1,108M$1.51——
Q4 2024$6,067M—$1,142M$1.59——
Q1 2025$7,061M—$2,005M$2.79$3.06 (derived)—
Q2 2025$6,974M+12.1%$1,829M$2.45$2.72vs $2.67
Q3 2025$6,351M+11.5%$1,170M$1.51$1.85vs $1.78
Q4 2025$6,595M+8.7%$1,219M$1.68$2.12vs $1.97
Q1 2026$7,597M+7.6%$1,754M$2.36$3.29vs $3.22
Q2 2026$7,404M+6.2%$1,899M$2.63$2.96vs $2.88

(Q1 2025 adjusted EPS derived as six-month 2025 adjusted of $5.78 less Q2 2025's $2.72; all other adjusted figures are from earn_cal and are confirmed by the 8-K reconciliation for the 2026 quarters.)

Four readings.

The revenue growth rate is decelerating in a straight line: +12.1%, +11.5%, +8.7%, +7.6%, +6.2%. On an underlying basis the six-month figure is 4%. That deceleration is the single most important fact in this file and it is the reason the multiple compressed.

The adjusted-to-GAAP bridge is large but fully itemised. The June quarter: GAAP EPS $2.63 plus $0.33 of adjustments equals $2.96. The $0.33 decomposes, per the 8-K reconciliation, into $267M of operating-income adjustments (restructuring $58M, contingent consideration $15M, McGriff integration and retention $47M, acquisition costs $2M, a legal matter $8M, and identified intangible amortisation of $137M), less $50M of other net benefit credits, less a $58M tax effect. A 12.5% adjustment is moderate by this programme's standards and the composition is disclosed line by line, which is more than most names in this batch manage.

But the six-month adjustment is 25.2% and that is not moderate. GAAP $4.99 to adjusted $6.25 is $1.26 per share, driven by $926M of operating-income adjustments — of which $651M is "noteworthy items" including the $425M Greensill charge and $275M is identified intangible amortisation. Intangible amortisation of roughly $550M a year is a permanent consequence of the acquisitive model and it will not go away; excluding it every year is a choice, and the reader should know that 16.8x FY2027 becomes roughly 20x on a basis that includes it. (Our arithmetic: $550M pre-tax ≈ $415M after tax ≈ $0.87 per share; $11.401 − $0.87 = $10.53; $191.45 / $10.53 = 18.2x.) We use the consensus adjusted basis throughout because that is what the estimates are struck on, and we state the alternative here so it is visible.

The beat pattern is consistent and small: +1.9%, +3.9%, +7.6%, +2.2%, +2.8%. Five consecutive beats, none large. This is a well-modelled company that guides carefully.

4. Cash flow, the balance sheet, and a debt figure the filing contradicts

Capex and free cash flow — checked first per standing instruction, and CLEAN.

YearOCFCapexFCFD&Acapex/D&A
FY2022$3,465M−$470M$2,995M$719M0.65x
FY2023$4,258M−$416M$3,842M$713M0.58x
FY2024$4,302M−$316M$3,986M$746M0.42x
FY2025$5,292M−$291M$5,001M$910M0.32x

Free cash flow is computed by subtraction and the arithmetic is correct in all four years. capexToDepreciationTTM of 0.362 sits below 1.0 — the signature that flagged corruption on COP, DUK, CSX, FCX and SO — but here it is explained rather than suspicious, and the filing supplies the explanation. D&A of $910M includes identified intangible amortisation of $275M in the first half of 2026 alone, i.e. roughly $550M a year, which is purchase-accounting amortisation from McGriff and two decades of bolt-ons, not consumption of physical assets. Stripping it, real depreciation is approximately $360M against $291M of capex — a ratio of 0.81x, which is normal for an asset-light advisory business. Corroboration: the 10-Q reports capital expenditures of $134M for the six months ended 2026-06-30 against $114M in the prior period, an annualised run rate of roughly $270M, consistent with FY2025's $291M. The 5.22% free-cash-flow yield stands as reported. No correction is applied.

The balance sheet, and the one figure where the vendor and the filing disagree:

Vendor bal_a 2025-12-3110-Q filed 2026-07-21
Cash and equivalents$2,687M—
Goodwill$24,337M—
Intangible assets$5,212M—
Total assets$58,710M—
Short-term debt / current portion$1,267M$1,267M ✓ (and $646M at 2026-06-30)
Long-term debt$18,320M$18,320M ✓ (and $18,891M at 2026-06-30)
Capital lease obligations$1,862M—
Total debt$21,449M$19,587M (= $18,320M + $1,267M)
Total equity$15,315M—
Minority interest$215M—

THE FILING WINS, AND THE DIFFERENCE IS 9.5%. The 10-Q's balance sheet reports long-term debt of $18,320M at 2025-12-31 and a current portion of long-term debt of $1,267M, which is $19,587M of debt. The vendor's totalDebt of $21,449M equals exactly $18,320M + $1,267M + $1,862M of capitalLeaseObligations — so lease liabilities are being added to a debt figure the balance sheet already reports in full. This is the T / TJX / GD / TT defect class, confirmed here by direct comparison to the filed statement.

Consequences, corrected:

VendorCorrected (filing)
Total debt$21,449M$19,587M
Net debt$18,762M$16,900M
Net debt / FY2025 EBITDA ($7,395M)2.54x2.29x
Vendor netDebtToEBITDATTM3.592x—
debtToEquityRatioTTM1.474x≈1.28x
Enterprise value$112,044M (vendor)≈$108,477M

The vendor's netDebtToEBITDATTM of 3.592x is the number a screen would use and it is roughly 57% above the leverage the filed balance sheet supports. All leverage figures in this dive use the filed debt.

Two further balance-sheet facts worth stating plainly. Goodwill of $24,337M plus intangibles of $5,212M is 48.4% of total assets against $15,100M of equity, so tangibleBookValuePerShareTTM is MINUS $27.85. That is not an error — it is what a company that has bought its growth for twenty years looks like, and it means there is no asset floor under this equity at all. And the goodwill step from $17,231M (2023) to $23,306M (2024) is McGriff, the acquisition that consumed $8,453M of cash in FY2024 against $652M in FY2025 and $181M across five deals in 2026.

5. Capital return and the corporate actions the payload missed

From the June-quarter release, verbatim: "The Company repurchased approximately 4.5 million shares of stock for $750 million in the second quarter of 2026. Through six months ended June 30, 2026, the Company has repurchased 8.7 million shares of stock for $1.5 billion. On July 8, the Board of Directors increased the quarterly dividend by 10% to $0.990 per share, with the third quarter dividend payable on August 14, 2026."

AmountAgainst $91.36B market cap
Buyback, six months to 2026-06-30$1,500M (8.7M shares)3.3% annualised
Dividend, new quarterly rate$0.990 (raised 10% on 2026-07-08)$3.96 annualised = 2.07%
Combined shareholder yield—≈5.4%
Vendor dividendPerShareTTM$3.69 (trailing)1.93% — understates the forward rate by 7.3%
Share count478M at 2026-06-30 vs 492M a year earlier−2.8% year on year

Two corporate actions absent from the vendor payload, both found in the 8-K record:

(1) The dividend increase itself. Declared 2026-07-08; dividendPerShareTTM of $3.69 and dividendYieldTTM of 1.93% are trailing figures that do not reflect it. The forward rate is $3.96 and the forward yield is 2.07%.

(2) A new $4.25 billion revolving credit facility. From Item 1.01 of the 8-K filed 2026-06-04: "On June 2, 2026, Marsh & McLennan Companies, Inc.... entered into a new Amended and Restated 5 Year Credit Agreement... The Credit Agreement provides for a multi-currency unsecured $4.25 billion five-year revolving credit facility... The New Facility expires in June 2031 and requires the Company to maintain certain coverage and leverage ratios which are tested quarterly." Item 1.02 records the simultaneous termination of the $3.5 billion facility it replaced. A $750 million upsizing of committed liquidity, and quarterly-tested leverage covenants that make the corrected debt figure in Section 4 operationally relevant rather than cosmetic.

(3) A rebrand and ticker change. The registrant is still "Marsh & McLennan Companies, Inc."; the market-facing identity is Marsh (NYSE: MRSH), the operating units are now Marsh Risk and Marsh Management Consulting, and the 8-K covers list NYSE Texas as an additional listing venue. The vendor profile carries the old company name and a description that still says "approximately 65,000 full-time professionals" against the 95,000 the company reports and the 95,000 in profile.fullTimeEmployees — the description block is internally inconsistent with its own record.

6. Valuation — priced in or room?

At $191.45 (market cap $91.36B, 478M shares at 2026-06-30):

TrailingFY2026EFY2027EFY2028E
Consensus adjusted EPS$9.75 (FY2025 actual, reconstructed)$10.448 (14)$11.401 (14)$12.481 (4)
EPS growth—+7.2%+9.1%+9.5%
P/E on adjusted—18.3x16.8x15.3x
P/E on GAAP trailing23.2x———
P/E including intangible amortisation (our arithmetic)—≈19.9x≈18.2x—
Consensus revenue$26,981M (FY2025)$28,582M (13)$29,994M (13)$31,556M (11)
Revenue growth—+5.9%+4.9%+5.2%
EV/EBITDA (corrected EV)≈18.9x———
Dividend yield (forward)2.07%———

Estimate coverage is good on the near years and thin beyond: 14 analysts on FY2026 and FY2027 EPS, only 4 on FY2028. The FY2028 row is used as a directional cross-check and nothing else. The FY2027 EPS range is $10.930 to $11.652, a 6.6% spread — the narrowest in this batch, which is what a well-modelled, low-beta fee business looks like and which means the fair-value work is almost entirely about the multiple rather than the estimate.

A useful calibration on the estimate block: the FY2025 epsAvg of $9.601 against the actual reconstructed adjusted EPS of $9.75 is a 1.5% understatement. The epsAvg row is trustworthy on this name, verified against an actual.

est.ebitdaAvg and est.ebitAvg are REJECTED — fixed-ratio fabrication. ebitdaAvg is exactly 27.24% of revenueAvg in FY2026, FY2027 and FY2028; ebitAvg is exactly 23.88% in every one. The ratio happens to sit close to the FY2025 actual EBITDA margin of 27.41%, which makes it look plausible and is precisely why it is dangerous. A number that is right by coincidence is still fabricated. All forward valuation uses epsAvg. Fifteenth consecutive name.

Peer set is, unusually, mostly correct. AJG, AON, BRO and WTW are the four genuine listed comparables and all four are present. CB and PGR are underwriters, not brokers; NTRS and ERIE are unrelated. The vendor payload does not carry peer earnings estimates, so no multiple comparison can be computed from this file, and none is asserted from memory.

6a. What today's price assumes (the inversion)

At $191.45 — 18.3x FY2026 consensus adjusted, 16.8x FY2027, 2.07% forward dividend yield — the price embeds:

6b. The return bridge (why the multiple moves)

Expected return over the next twelve months decomposes as: adjusted EPS growth (+9.1%, from FY2026E $10.448 to FY2027E $11.401) + multiple drift (MODEST EXPANSION; our base of $208 is 18.2x FY2027E against 16.8x today, +8.6%) + shareholder yield (+5.4%: 3.3% buyback and 2.07% dividend).

This is the one name in this batch whose base case requires multiple expansion, and we say so as the fragile leg rather than burying it. Our $208 assumes the market pays 18.2x FY2027 adjusted earnings rather than today's 16.8x — an 8% re-rating. The argument for it is that the three causes of the de-rating are identifiable and two of them are finite: the Greensill charge is taken, the McGriff integration costs ($89M in the first half, against $114M in the prior-year half) are declining, and only the property-and-casualty rate cycle is genuinely open-ended. The argument against it is that a business growing 3% underlying in two-thirds of its revenue does not deserve 20x, and that the market may be right.

Note the composition: roughly 9 points of earnings growth, roughly 5 points of shareholder yield, and roughly 8 points of hoped-for re-rating. If the re-rating does not come, the return is still 14% — earnings plus yield — and the stock is worth roughly $192, which is spot. That asymmetry is the actual case for Watch: the downside case is "you earn the coupon and nothing else," which is not a loss.

6c. Variant perception (where we differ, what would surprise)

Synthos fair values

All three anchors are multiples of the FY2027 consensus adjusted EPS distribution (mean $11.401, low $10.930, high $11.652, 14 analysts), each cross-checked against FY2026 and against the shareholder yield.

Base is 8.6% above spot; asymmetry 1.69:1 (17.5% down, 29.5% up), plus a 5.4% shareholder yield in every scenario. That is a genuinely reasonable proposition and it is not a Buy: this programme requires roughly 2:1 with named triggers for a Tactical position, and 1.69:1 into a stock that has risen 14% in three months and sits 10.8% above its 50-day average is not it. Watch, with the trigger prices stated in Section 9.

7. Knowledge base — a genuinely empty lane

Raw hits: 13. After a case-sensitive re-run: 4. Entity matches: ZERO. Used: 0. Discarded: 13.

The sweep ran entity terms MRSH, MMC, Marsh, Marsh & McLennan, Marsh McLennan, Aon, AON, Willis Towers Watson, WTW and AJG, plus free text on marsh, insurance broking, P&C insurance, reinsurance and Mercer, across the full distilled claim store.

Not one claim in the store carries Marsh & McLennan — or Aon, or Willis Towers Watson, or Arthur J. Gallagher, or Brown & Brown — as an entity. Every one of the four case-sensitive survivors is a free-text match with an empty entity list:

> 2026-03-14 · neutral · conviction 70 · horizon: principle · entities: (none) · channel: we_study_billionaires

> "Long-track-record investors must evolve within their wheelhouse (Davis into reinsurance/international; Buffett from cigar-butts to quality) to keep succeeding."

> 2025-01-23 · neutral · conviction 75 · horizon: principle · entities: (none) · channel: no_priors

> "Regulated admitted insurers barred from pricing in reinsurance costs and forward-looking risk models will exit high-risk markets — market failure until rules change."

> 2026-04-10 · neutral · conviction 80 · horizon: principle · entities: (none) · channel: odd_lots

> "The 12 International Group P&I clubs insure ~90% of oceangoing tonnage and jointly buy the world's largest reinsurance policy, giving purchasing power for cheap high-limit cover (~$3B) despite being competitors."

> 2026-06-01 · bullish · conviction 68 · horizon: thesis · entities: (none) · channel: empire · speaker_role: independent

> "Re, an on-chain reinsurance business, uses stablecoin markets as its capital function, paying real 12-14% yield from a genuinely productive business — 5-7x leverage on treasuries. In portfolio."

Three of those four are genuinely interesting about the reinsurance market and none is usable. The odd_lots claim is the most relevant in spirit — it describes buyers organising collectively to obtain cheap high-limit cover, which is a description of what happens to intermediation economics when clients get large and sophisticated — but it names no company, carries no entity, and inferring a Marsh implication from it would be exactly the kind of manufactured conviction this programme exists to avoid.

The remaining nine raw hits are pure substring noise: all_in (2025-10-02) on a "Marshall Plan for America"; three biotech_hangout hits; two paul_glimcher hits; and others where "marsh" appears inside another word. No homograph collision of the CEG or PWR class occurred — the failure here is simpler and cleaner: the store has nothing on this industry.

Conclusion. Breadth 0, claim count 0, net conviction NONE. The lane is EMPTY and is reported as empty rather than padded, consistent with the treatment of CSX and JCI. Conviction contributes exactly zero to this verdict.

One finding worth logging beyond this dive: the entire listed insurance-brokerage industry — Marsh, Aon, Willis Towers Watson, Arthur J. Gallagher and Brown & Brown, a group with a combined market capitalisation above $280 billion — has zero entity coverage in this knowledge base. That is a sector-level coverage gap rather than a defect, and it should be raised with the ingest side rather than worked around at the dive level.

8. Data integrity — what we rejected and why

Seven findings. Two are among the most concrete in this batch and both were settled by reading the filed statements.

1. totalDebt is 9.5% ABOVE the filed figure — lease obligations added to debt already reported. THE FILING WINS. The 10-Q balance sheet reports long-term debt of $18,320M and a current portion of $1,267M at 2025-12-31, i.e. $19,587M. The vendor's totalDebt of $21,449M equals exactly that plus $1,862M of capitalLeaseObligations. Corrected: net debt $16,900M rather than $18,762M; net debt / FY2025 EBITDA 2.29x rather than 2.54x; enterprise value approximately $108,477M rather than $112,044M; debtToEquityRatioTTM approximately 1.28x rather than 1.474x. The vendor's netDebtToEBITDATTM of 3.592x is not used anywhere in this dive. Corroboration from the following period: the 10-Q reports long-term debt of $18,891M and a current portion of $646M at 2026-06-30. This is the T / TJX / GD / TT defect class.

2. seg_prod FY2025 is the worst segment failure recorded in this programme — REJECTED outright. The four FY2025 lines are Marsh Insurance Group $31,492M, Mercer Consulting Group $6,190M, Oliver Wyman Group Consulting Group $3,604M and Guy Carpenter Reinsurance Group $2,635M. They sum to $43,921M against total company revenue of $26,981M — 162.8%. The Marsh Insurance Group line ALONE is $31,492M, which is 116.7% of the entire company's revenue. The true figure is derivable by subtraction — total revenue $26,981M less Mercer, Oliver Wyman and Guy Carpenter gives approximately $14,552M, so the reported line is 2.16x the actual. This is not a rounding issue or an aggregate/component double-count; it is a different quantity entirely, most plausibly premiums placed rather than revenue. Note that the FY2024 row uses the correct two-segment structure and ties: Risk and Insurance Services $15,395M plus Consulting $9,133M equals $24,528M against $24,458M of revenue, 100.3%. The block was correct until the most recent year — i.e. the year a reader would use. All segment figures in this dive come from the 8-K and 10-Q.

3. seg_geo FY2025 ties and is ACCEPTED. United States $13,344M + United Kingdom $3,816M + Other Geographic Areas $9,899M = $27,059M against $26,981M of revenue, 100.3%. Correctly labelled, no double-counting. Recorded as a clean result. Note that the block collapsed from five regions (through FY2022) to three (from FY2023), so Asia Pacific and Europe are no longer separately visible — a loss of granularity, not an error.

4. Corporate actions absent from the payload — THREE. (a) The dividend increase of 2026-07-08, raising the quarterly rate 10% to $0.990 ($3.96 annualised, 2.07% forward yield); dividendPerShareTTM of $3.69 and dividendYieldTTM of 1.93% are trailing and understate the forward rate by 7.3%. (b) A new $4.25 billion five-year revolving credit facility entered 2026-06-02, expiring June 2031, replacing a terminated $3.5 billion facility, with coverage and leverage ratios tested quarterly. (c) The rebrand and ticker change — the market-facing identity is now Marsh (NYSE: MRSH), the operating units are Marsh Risk and Marsh Management Consulting, and the 8-K covers add NYSE Texas as a listing venue. None appears in any vendor field.

5. est.ebitdaAvg and est.ebitAvg carry the fixed-ratio fabrication signature — REJECTED, with a warning. ebitdaAvg is exactly 27.24% of revenueAvg and ebitAvg exactly 23.88% in FY2026, FY2027 and FY2028. What makes this instance instructive is that the fabricated ratio is very close to the truth — FY2025 actual EBITDA of $7,395M on $26,981M is a 27.41% margin — so the numbers look right. A figure that is accidentally accurate is still fabricated and is still rejected. All forward valuation uses epsAvg, which was independently verified: FY2025 epsAvg of $9.601 against a reconstructed actual adjusted EPS of $9.75, a 1.5% understatement.

6. returnOnEquityTTM is present in km_ttm (0.2653) and ABSENT from ratios_ttm — fifteenth name running. The km_ttm figure of 26.53% is used. Note that on a company with negative tangible equity, return on equity is a partly artificial measure; returnOnInvestedCapitalTTM of 11.40% and returnOnCapitalEmployedTTM of 15.77% are the more informative figures and both are used in preference where relevant.

7. profile is internally inconsistent and stale. fullTimeEmployees reports 95,000, matching the company's "more than 95,000 colleagues" — but the description field in the same block states "employing approximately 65,000 full-time professionals." The two figures are 46% apart within one record. The description also carries the pre-rebrand company name alongside the new ticker. The profile block is used only for sector, industry, beta and chief executive, all of which check out (John Q. Doyle; Insurance — Brokers; beta 0.603).

Enterprise value — residual quantified and explained. enterpriseValueTTM of $112,044M against a rebuild on vendor inputs of market cap $91,362M + total debt $21,449M − cash $2,687M + minority interest $215M = $110,339M, a 1.5% residual. On corrected (filed) debt the rebuild is $108,477M, and the vendor figure is 3.3% high. Within the 15% withholding threshold either way; the corrected figure is used.

Share count — CHECKED AND MATCHED. The 10-Q reports 478 million shares outstanding at 2026-06-30 (down from 492 million a year earlier) against 477.2 million implied by market capitalisation ÷ price. A 0.2% match.

Quote versus technical blocks — minor 52-week disagreement. quote.yearHigh/yearLow report $213.80 / $156.60; tech.hi52/lo52 report $211.21 / $157.32. Differences of 1.2% and 0.5%. tech is used throughout.

Non-equity tripwire — checked and passed. Common stock, NYSE-listed (and now NYSE Texas), per the 8-K cover pages. Price $191.45; beta 0.603; volume 2.01M shares (~$385M of turnover); 52-week band $157.32-$211.21, a 34% range; dividend variable and just raised. This is common equity.

9. Technicals, insiders, and the verdict

Technicals

Today's move. Marsh closed DOWN 0.72% at $191.45 from $192.835, opening at $189.15 and trading $188.83 to $192.24 on 2.01M shares — a quiet down day in a strong tape. No company-specific filing is dated 2026-08-04; the last event was the 2026-07-21 results.

The honest read on the entry: it is not a discount. The valuation case was made at $157 in the last twelve months and the stock has already recovered 22% of it. A Watch designation with a stated trigger price is the correct expression of that, and the trigger is in the flip conditions below.

Insiders — three open-market sales, one of them a complete exit

DatePersonRoleTypeSharesPriceHeld after
2026-07-29Paul BeswickSVP, Chief Information OfficerS-Sale713$198.0016,376
2026-07-28Martin SouthChief Client OfficerS-Sale7,100$189.9916,631.7
2026-07-23Lloyd M. YatesDirectorS-Sale12,000$175.6512,000.75
2026-05-15Morton O. SchapiroDirectorJ-Other (RSUs)495.08$089,095.6
2026-05-15Morton O. SchapiroDirectorA-Award256.85$088,600.5
2026-05-15Deborah C. HopkinsDirectorJ-Other (RSUs)111.82$020,064.7
2026-05-15Tamara IngramDirectorJ-Other (RSUs)43.11$07,736.5
2026-05-15Lloyd M. YatesDirectorJ-Other (RSUs)60.27$011,041.2

The reading. Five of the eight transactions are mechanical director restricted-stock-unit accruals dated 2026-05-15 and carry no signal. Three are open-market sales, all executed in the eight days between 2026-07-23 and 2026-07-31 — i.e. in the window bracketing the 2026-07-21 earnings release — and there is not one purchase in the file.

The largest and most notable is Lloyd M. Yates, a director, who sold 12,000 shares at $175.65 on 2026-07-23 and was left holding 12,000.75 in his direct account — approximately half his direct position, executed two days after results at a price 8.3% below where the stock closed twelve days later. Note that the same person accrued 60.27 restricted units on 2026-05-15 and held 11,041 units then, so the reported "12,000.75 after" figure most plausibly reflects a different account or instrument class; we report the vendor's numbers as given and flag the ambiguity rather than reconciling it by assumption.

Martin South, Chief Client Officer, sold 7,100 shares at $189.99 on 2026-07-28, a 29.9% reduction. Paul Beswick, Chief Information Officer, sold 713 at $198.00, a 4.2% reduction, at the highest price in the table.

Total open-market disposals: 19,813 shares, approximately $3.6 million. The amounts are small in absolute terms; the pattern — three officers and directors selling in the week after results, no purchases, and the largest sale a director halving a holding — is mildly negative and is reported as mildly negative.

10. Verdict, kill-criteria and flip conditions

Watch.

The arithmetic. At $191.45 Marsh trades at 18.3x the FY2026 consensus adjusted EPS of $10.448 and 16.8x FY2027's $11.401 — approximately 18.2x on a basis that includes the roughly $550 million a year of intangible amortisation consensus excludes. Our base fair value of $208 is 8.6% above spot with 1.69:1 asymmetry ($158 bear, $248 bull), plus a 5.4% shareholder yield. The base case requires an 8% multiple expansion and we name that as the fragile leg.

What is genuinely good. Adjusted EPS +9% in the quarter and +8% for the half; revenue +6% and +5% underlying; a 30.5% adjusted operating margin on the half; 26.5% return on equity; 0.603 beta; recurring fee revenue from 130 countries; capital intensity of 1.1% of revenue with free cash flow of $5,001M; $1.5 billion of stock repurchased in six months against 478 million shares outstanding, down 2.8% year on year; and a dividend raised 10% on 2026-07-08 to $0.990 a quarter. And a $425 million Greensill charge — a five-year overhang — has now been quantified and partly settled.

What keeps it at Watch rather than Buy.

First, the asymmetry does not clear the bar. 1.69:1 against a standing requirement of roughly 2:1 with named triggers. The gap is not large and it is not close enough to argue about.

Second, the entry is into a 14% three-month rally, with the price 10.8% above its 50-day average and RSI at 66.5. The valuation case was available at $157 within the last twelve months and 22% of it has already been collected by somebody else.

Third, underlying growth in the core is 3%. Risk and Insurance Services is roughly two-thirds of revenue and grew 3% underlying in both the quarter and the half, against Consulting's 8%. Nothing in the filings distinguishes a property-and-casualty rate trough from commission compression, and the difference between those two readings is most of the fair-value range.

Fourth, the adjusted margin is not expanding despite Thrive. 29.3% against 29.5%; 30.5% against 30.7%. A $500 million programme targeting $400 million of savings that produces no visible margin gain in its fourth quarter of operation is a programme to watch rather than to underwrite.

Fifth, the knowledge base is empty and the insiders are selling. Zero name-level claims; three open-market sales in the week after results and no purchases.

Pre-registered KILL criteria — what would take this to Hold or Avoid:

Pre-registered UPGRADE conditions — what would take this to Buy — Tactical:

Where MRSH fits in the Synthos Framework Portfolio. No position today; a named limit order. This is a high-quality, low-beta, cash-generative franchise trading at the bottom of its own multiple range for reasons that are two-thirds finite, and the correct response is a 1.5% starter position on a pullback toward $170, scaling to 3% on confirmation that underlying growth in the core has troughed. Buying it 10.8% above its 50-day average after a 14% run, when the same case was available 22% lower this year, is paying for someone else's patience. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $191.45, with the fair-value anchors, kill criteria and the $170 trigger all gradeable.

Single biggest risk: underlying growth in Risk and Insurance Services at 3%. Two-thirds of the company grew 3% on an underlying basis in both the June quarter and the six months, while the remaining third grew 8%. If that is the property-and-casualty rate cycle, it turns and the multiple recovers and the base case is conservative. If it is commission compression or share loss to specialists, then a business growing 3% in its core does not deserve 18x and the fair value is nearer the bear case than the base. The filings do not distinguish the two, the analyst community has evidently given up trying — 22 Holds out of 34 — and the October print is the next piece of evidence. That is exactly what a Watch is for.

Provenance & disclosures