Financebotresearch desk研究台

MEDP › analyze

MEDP · Analyze

WATCH Healthcare

Healthcare / Diagnostics & Research (Contract Research Organization) · United States · ~6,500 employees Price $583.16 · Market cap $16.28B · 52wk $373.00 – $677.90 (a peak-to-trough range of −45%, then +82% off the low — this stock re-rates violently on a single metric, see §3)

One line: A genuinely elite niche compounder — 20%+ growth, 58% ROIC, net cash, buybacks shrinking the float — that got its own management sued for allegedly gaming the exact metric (book-to-bill / cancellations) this thesis has to lean on, and hasn't yet earned back the multiple it lost. Good business, bad time to pay up for it.

0. Knowledge check

python .mcp/kb.py find MEDP → no prior notes; python .mcp/kb.py find CRO → no field sweep on contract research organizations. This is the agency's first look at MEDP and at the CRO field. Relevant Playbook items checked before trusting any vendor figure:

  • [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] — applies directly. Yahoo's forwardPE of 30.2x does not disclose which fiscal year it prices. Management's own FY26 EPS guide ($17.25–$17.95, midpoint $17.60) gives a fwd P/E of 33.1x at spot — the more trustworthy anchor, and it is what this report uses throughout. StockStory's independent 32x fwd read (below, §3) corroborates this over the vendor figure.
  • [[pattern-incremental-margin-below-average-locks-in-decay]] — applied in §1: it correctly flags that MEDP's 2023→2024 margin expansion is over, not merely slowing.
  • No pitfall exists yet for the specific pattern this name surfaces — a CRO's own management can be sued for the book-to-bill/backlog guide itself, not just for vendors mis-stating it. Worth a note after this file lands (outside this report's scope-lock).

1. Fundamentals

Free cash flow and growth

FY2022 FY2023 FY2024 FY2025
Revenue $1.46B $1.89B $2.11B $2.53B
Op. income $278.7M $336.8M $446.9M $534.9M
Net income $245.4M $282.8M $404.4M $451.1M
OCF $388.1M $433.4M $608.8M $713.2M
FCF $351.2M $396.7M $572.3M $681.9M
Diluted shares 33.67M 31.84M 32.01M 29.53M
FCF/share $10.43 $12.46 $17.88 $23.09

3-year CAGRs (2022→2025, the only window fin.py and roic.ai's free tier can both reach — flagged as a data gap, not a 5-8yr read): Revenue +20.1%, net income +22.5%, FCF +24.8%, FCF/share +30.8% (buybacks add ~6pts/yr on top of aggregate FCF growth). Genuinely fast, genuinely profitable growth — not bought with debt or dilution.

Capital allocation of FCF (FY2025): $917.4M returned via buybacks against $681.9M of FCF — 134% of the year's cash generation, funded partly by drawing cash down ($669M→$497M). Zero dividend, zero net debt issuance, capex is a rounding error (1.2% of revenue — this is a people-and-office business, not a capital-intensive one). This is about as aggressive and as clean a buyback program as the framework sees: no M&A, no debt-funded repurchases, all excess cash to shareholders through the share count. Diluted shares fell −7.7% in FY2025 alone.

Debt-to-assets: 5.75%. Debt $113.6M vs. cash $497.0M — net cash ~$383M. Current ratio 0.73x looks tight but is structural for a CRO: net DSO is negative 59.6 days (clients routinely prepay), so working capital is a funding source, not a drag.

The incremental-margin check ([[pattern-incremental-margin-below-average-locks-in-decay]] applied): FY2023→FY2024, incremental operating margin was ~50.2% against a 17.8% average — real operating leverage, margin expanded 17.8%→21.2%. FY2024→FY2025, incremental margin fell to 20.9% against a now-higher 21.1% average — essentially at parity. The identity says the expansion phase is over: FY2025 margin was flat (21.19%→21.14%), and absent a new leverage source, FY2026 margin should also be roughly flat. This is a volume-growth story now, not a margin-expansion story — worth knowing before crediting the bull case with further margin upside it hasn't earned since 2024.

§1 scorecard: strong. Fast, clean, per-share-accretive growth on a debt-free balance sheet. The only flags are the ones above (short lookback window, margin expansion already spent) and the top-line deceleration embedded in guidance itself — FY2026 revenue is guided to 10.9–14%, roughly half the recent pace (see §3 for why).

2. Moat

Quantitative base. ROIC (roic.ai): 45.9% (FY2024) → 58.4% (FY2025), rising and exceptional — for comparison, IQVIA (the largest full-service CRO) runs ROIC ~8.3%. Gross margin ~30-31%, stable. This is a real, high-return niche, not a commodity-services roll-up multiple arbitrage.

Revenue-stream map. Single stream — full-service outsourced clinical trial management (Phase 1-3), sold almost entirely to small and mid-cap biotech, the segment the large CROs (IQVIA, ICON, Thermo Fisher/PPD) are not built to serve well: those firms are organized around large multi-country programs for big pharma. Medpace's pitch to a 40-person biotech with no internal clinical-operations function is a single dedicated therapeutic-area team end-to-end. That is a real switching cost — Morningstar rates it a narrow moat built on intangible therapeutic expertise and switching costs — and the ROIC gap to IQVIA is the receipt.

Adversarial stress-test. Could a well-funded rival take this? Two live vectors say the moat is real but not effortless: - Management openly admitted a declining win rate in 2025 and had to make unspecified "competitive changes" (declined to detail them on the Q2 2026 call) to recover it — a moat under quiet, active price/positioning pressure does not describe an unassailable position. - Customer concentration is rising, not falling: top-5 clients are now 31% and top-10 40% of LTM revenue, driven by a handful of large cardiometabolic (GLP-1-adjacent) programs. A single large-program loss or completion is a bigger air pocket than it was two years ago.

Disruption forecast (5-10yr). The whole CRO industry faces the same slow-moving threat: AI- assisted trial design, synthetic control arms, and decentralized/remote trial infrastructure could compress billable clinical-operations hours over the decade. Not an near-term threat to Medpace specifically, but not a moat defense either — nobody in the field has answered it yet.

Evergreen assessment. Qualifies on the durability test (real switching costs, high ROIC, founder-CEO since 1992, low capital intensity) but fails the "hold through a 40% drawdown without re-underwriting" test as written today — the name already delivered a 45% drawdown in early 2026, and the reason was a disclosure controversy about the business's own core metric, not a cyclical funding air-pocket the market mispriced. That is a different, harder-to-underwrite risk than "biotech funding is cyclical" — see §3.

3. Sentiment — the book-to-bill / litigation timeline (the central risk)

This is the finding the task explicitly asked to watch for, and it is not hypothetical here — it already happened once, in public, with a lawsuit attached.

Date Event
Apr 2025 – Feb 2026 Class period. Management repeatedly told investors cancellations were "well behaved" / isolated, and guided/reiterated a 1.15x book-to-bill expectation for Q4 2025.
Feb 9, 2026 Q4 2025 actual book-to-bill came in at 1.04x; CEO Troendle called cancellations "the highest in over a year." Stock fell ~15.9% in one session.
May 2026 Multiple securities-fraud firms file Durbin v. Medpace Holdings (S.D. Ohio, No. 1:26-cv-00346), alleging Medpace oversold the book-to-bill guide and concealed a worsening cancellation trend during the class period. Lead-plaintiff deadline June 8, 2026.
Q1 2026 Book-to-bill fell further to 0.88x — backlog shrank net of cancellations that quarter. Stock printed the 52wk low of $373.
Jul 23, 2026 (Q2 2026 print) Sharp reversal: net new bookings a record $795.7M, net book-to-bill 1.13x (+28.2% YoY), cancellations "well behaved" again, FY2026 guidance raised (revenue, EBITDA, EPS all up). Stock +17% in one day; 52wk high $677.90 followed.
Aug 12–24, 2026 CEO Troendle sells ~$40M+ of stock across six separate Form 4s, described as executed "pursuant to a limit order placed during an open window period."
Sep 10, 2026 (today) Litigation status: pending, no motion-to-dismiss ruling found in public search as of this report.

Reading it straight, not just the headline quarter. Backlog as of June 30, 2026 was $3.01B, +4.9% YoY — modest, despite the record bookings quarter, because the backlog conversion (burn) rate structurally rose from a historical ~18% to ~24% of beginning backlog per quarter. On the Q2 call, management attributed most of this to tightening what counts as backlog (more conservative gating around interim-analysis decision points) after the 2025 cancellation wave — i.e., the backlog dollar figure is arguably higher quality now, but it also mechanically converts to revenue faster, so the +4.9% YoY growth print understates underlying momentum somewhat. That is a genuinely nuanced, not purely bearish, explanation — but note that it is the company's own explanation for the same metric it is being sued over having explained away before. The CEO has also now explicitly stopped forecasting cancellations at all ("completely beyond my ability to even guess") — more honest, but it means investors have less forward visibility on the metric that broke the stock than they had a year ago, not more.

Biotech funding backdrop (WebSearch, Sep 2026): broadly improving — 2026 H1 VC funding into biotech was the highest first-half total since 2022, and Medpace's own RFP volume and funded- client mix corroborate a real recovery, not just a bookings-timing artifact. But multiple industry sources flag a widening gap between well-capitalized and smaller biotechs — precisely Medpace's core client base — so the sector tailwind is real but uneven, and the small end of it (where Medpace lives) is the part most exposed if funding cools again.

Insider activity. The August 2026 sales came after the scandal, during an open window, and are framed as limit-order executions rather than a pre-scheduled 10b5-1 program set months in advance — a weaker governance signal than a genuine 10b5-1 plan, though not proof of anything by itself. Given the pending fraud suit alleges leadership knew about deteriorating conditions before disclosing them, continued heavy CEO selling into a rallying stock is a fact pattern worth tracking, not dismissing — see [[pitfall-market-sell-to-cover-is-coded-S-not-F]]-adjacent caution: confirm the code and plan type before treating any single Form 4 as routine.

Independent third-party read (StockStory, Sep 9, 2026): names MEDP one of "2 healthcare stocks with competitive advantages," citing 17.2% 2yr organic growth, 29.6% EPS growth (buyback-boosted), and a 5-point FCF-margin gain over five years — a genuinely positive outside read, at a cited 32x forward P/E, corroborating this report's 33.1x calculation off the company's own guide.

4. Valuation

Graham's Intrinsic Value — flagged as non-diagnostic here. √(22.5 × $17.00 × $15.54) ≈ $77, a small fraction of the $583 price. This is a model artifact, not a signal: BVPS is mechanically depressed by $917M of FY2025 buybacks eating equity down to $459M, and tangible book value is already negative (equity $459M − goodwill $662M ≈ −$203M, or roughly −$7/share). Graham requires a real book-value anchor; a hyper-aggressive-buyback, asset-light compounder doesn't have one. Per the framework's own rule for non-dividend, high-growth names, weight this near zero and lean on FCF/earnings multiples instead.

DYT / DDM: not applicable — no dividend (payout ratio 0%).

FCF/earnings multiple approach. FY2025 P/FCF was 24.3x on a P/FCF band of 10.8x–27.1x over the trailing year — the multiple itself round-tripped by more than 2x as the stock fell to $373 and recovered to $678. That range is the risk profile in one number: this name re-rates its multiple violently around single-quarter book-to-bill prints. On the company's own FY2026 EPS guide (midpoint $17.60), spot $583.16 is 33.1x forward — above its own FY2024 average forward multiple (~28x) and near the top of the FY2025 range, despite the unresolved litigation overhang.

Bogle's Expected Return (qualitative, stated assumptions): 0% dividend yield + an assumed ~15-18% sustainable EPS growth (guide-consistent, buyback-assisted) minus an assumed multiple mean-reversion from ~33x toward a ~26-28x normalized band (a mid-single-digit annual drag) nets to roughly low-double-digit expected annual return from here — a decent but unexciting number for a name carrying open litigation risk, not the "get in before it re-rates" setup a 20%+ grower would usually offer.

Fair value range: $460–560 (26–32x FY2026 guide-midpoint EPS $17.60), reflecting real quality but a valuation that should carry a discount for unresolved fraud litigation and demonstrated extreme volatility, not the premium multiple a clean compounder with this ROIC would otherwise command. Entry $422–475 (24–27x fwd) — meaningfully below spot, i.e., no urgency to chase it here. Trim 30x fwd (~$528) — current spot is already above this ceiling, which is the valuation case for WATCH rather than BUY: on the numbers alone this is a name to want at a lower price or after litigation clarity, not at $583.

5. Tension worth naming explicitly

Fundamentals + Moat argue for a strong BUY-tier business (20%+ growth, 58% ROIC, net cash, real niche switching costs). Sentiment + Valuation argue for caution: an unresolved securities-fraud suit about the exact metric the moat case leans on, continued heavy insider selling into the recovery, demonstrated 82%-off-the-low volatility, and a price that is already through the trim ceiling on the company's own guidance. Weighting: for a niche services compounder where trust in management's own forward commentary is the thing in question, Sentiment and Valuation should outweigh Fundamentals and Moat until the litigation resolves or the price gives back the premium — hence WATCH, not BUY, despite the business quality.

6. Verdict

§1 health bar: Strong — clean, fast, per-share-accretive growth on a debt-free balance sheet. No red flags in the fundamentals themselves; the only caveats are a short (~3yr) verifiable CAGR window and a spent margin-expansion phase (§1).

Buy zone or watch: WATCH. Entry $422–475 (24-27x fwd), current spot $583 sits ~23% above the top of that zone and above the proposed trim ceiling. This is not a name to initiate today regardless of how the litigation resolves — pay the premium later, on a pullback or on litigation clarity, not into a post-beat rally.

Does it beat the weakest Evergreen members (LHX WATCH 5.5 / PCTY-IBM 5.5)? Yes, on business quality — earns a sleeve line. LHX has no asset floor either (tangible book −$30.6/share) but pairs that with decelerating growth (9.5%→7.1%→2.3%) and an unresolved board-governance question around a for-cause firing; PCTY and IBM are both mid-single-digit-growth, margin-defense stories with softening retention/dividend metrics. MEDP's underlying economics — 20%+ growth (even at the guided-down 10.9-14%, still 4-6x LHX's rate), 58% ROIC vs. LHX's implied-lower and IQVIA's 8.3%, genuine net cash — are simply better businesses than the current floor of this sleeve. What keeps it from scoring meaningfully above that floor is the fresh, name-specific overhang the other three don't carry: an active fraud suit about the company's own disclosure of the metric this report had to center, plus continued CEO selling during it.

Proposed sleeve + conviction: 🏛 Evergreen Compounders, conviction 6.0, verdict WATCH. Above the sleeve floor on quality, capped below a stronger score by litigation/governance risk that is specific to this name, not the sector. Revisit at the Q3 2026 print (watch net book-to-bill and whether cancellation language repeats "well behaved") or on any motion-to-dismiss ruling in Durbin v. Medpace Holdings, whichever comes first.

Biggest risk: the book-to-bill/backlog metric is not just cyclically noisy here — management was sued for allegedly misrepresenting it once already (Q4 2025 disclosure, −15.9% one day, stock to a 52wk low four months later), the company has since stopped forecasting the metric at all, and the CEO kept selling stock through the recovery rally that followed. The fundamentals are real; the disclosure trust is not yet rebuilt, and the price does not yet reflect that gap.

Sources