# Compounder Screen - August 2026 > [!warning] This note goes stale > Everything here is priced as of **21 August 2026**. The frameworks in [[Quality Compounders]] and [[Acquisition-Driven Compounders MOC]] are evergreen; this is their output on one particular day. Re-run it, don't reread it. Nothing here is advice — it is a screen result and the questions the screen cannot answer. ## The regime The setup is unusual, and worth stating before any name. - **The S&P 500 trades at 29.7× trailing / 20.0× forward.** US 10-year at 4.71%. Riksbank at 1.75% and openly restating the chance of a hike this year; the ECB *raised* to 2.25% in June 2026. - **Quality just had its worst run in sixty years.** Eastspring puts the quality factor's H1 2026 drawdown at a six-decade low, with the Fama-French profitability factor at levels last seen in the dot-com bubble. Small caps +23.9% in H1 against +10.2% for the S&P; the Philadelphia Semiconductor Index doubled; Micron, Intel and AMD alone were 34% of the index's gain. - **The Nordic serial-acquirer cohort now trades below its own five-year median** — ~14× LTM EV/EBITDA against a five-year median of ~16×, and ~32× P/E against ~38× (Redeye, Q4 2025 update). Median organic growth ~1%. Median net debt/EBITDA ~2.2×, flat. **No covenant distress, no rescue equity, anywhere in the cohort.** - **Cohort M&A volume hit a seven-year low in 2025.** Nordic private mid-market assets still clear around **5.5× EBITDA** (Dealsuite). Listed acquirers trade at 14×. > [!important] The engine is still open > A 14× listed multiple buying 5.5× private assets is the entire arithmetic of [[Acquisition Multiples and Deal Structure]]. That spread has narrowed since 2021 but it has not closed. What has slowed is *deployment*, not *opportunity*. ## The sorting that actually happened This is not a sector de-rating. It is a re-sort, and the axis is AI. | Down hard, 12 months | | Up hard, 12 months | | | --- | --- | --- | --- | | Lumine | −55% | Nordson | +59% | | Addnode | −49% | Amphenol | +45% | | Topicus | −43% | IDEX | +43% | | Judges Scientific | −40% | Diploma | +37% | | Tyler Technologies | −39% | NCAB | +34% | | Vitec | −35% | DCC | +33% | | Constellation Software | −31% | AMETEK | +30% | | Lifco / IMCD / Röko | −24% | OEM International | +30% | | Roper / Brown & Brown / Watsco | −24% | HEICO | +15% | **Everything falling is vertical-market software or slow-organic distribution. Everything rising touches AI datacentres, semiconductors or defence.** The two lists are the same bet seen from opposite ends: whether AI commoditises niche software. ## Where the screen points Run [[The Compounder Screener]] against current filings. Passing most lines, in rough order of how cleanly: **Lifco** — organic +4.7% in Q2 and re-accelerating, net debt/EBITDA 1.8× (1.2× interest-bearing, against a 2–3× target — *under*-levered), ROCE 20.4%, EBITA margin 23.1% and rising, 17 deals in 2025 and 5 in H1 2026, no CEO change, no equity issued, no controversy. Down 24% on the year with nothing operational behind it. Still 19.9× EV/EBITDA — passing the quality screen does not make it cheap. **Indutrade** — organic +5%, order intake +8%, net debt/EBITDA 1.7×, cash conversion 140%, record EPS +23%, margin 14.7% from 13.7%, 9 deals. Flat on the year. The [[Case File - Indutrade]] pattern is still intact eighteen years after IPO. **Röko** — organic +5% H1, EBITA margin 22% stable, 4 deals adding SEK 1.25bn of sales, **and the only company in the entire European cohort that publishes what it pays: average EBITA multiples "below 8×."** Leverage 2.6× is marginally over your 2.5× line. Founder Fredrik Karlsson handed over to CFO Johan Bladh in April 2026 and stayed as Deputy CEO and chair of the investment committee — the internal-candidate succession [[Ownership Insiders and Succession]] rates highest. Trading below its 2025 IPO price, 36% off its high. Least proven name on this list by a distance. **Halma** — organic +16% constant currency in FY26 (roughly half of that is photonics premium expansion, so read it nearer 8% underlying), net debt/EBITDA 1.16×, 5 deals for £447m, no leadership change, FY27 guided to low-double-digit organic. 27× forward. **Diploma** — the best operations in the dataset: organic +15% for nine months, guidance upgraded twice, margin upgraded to ~26.5%, 15 deals and ~£300m deployed in twelve months. Also the most expensive thing in Europe at 31.9× EV/EBITDA, up 37%, **and it does not disclose net debt/EBITDA** — which is on your own yellow-flag list. > [!tip] The widest gap between quality and price > **Vitec Software.** Organic recurring growth 5–6% with a 1–1.5% pricing tailwind, EBITA margin 28–29%, revenue +15%, no CEO change, no profit warning, no equity raise, no strategy change, no controversy — and the stock is down 35% to **8.1× EV/EBITDA and 17.4× forward earnings.** A vertical-market-software compounder at single-digit EV/EBITDA. There is no operational trigger. The entire move is the AI question. ## What the screen kills - **Brown & Brown** — organic **−0.7%**, a $9.8bn acquisition funded with **$4bn of new equity**. Two of the loudest red flags in [[Red Flags in Serial Acquirers]] in one company. - **Addnode** — organic −5% adjusted, −11% unadjusted; EBITA −38%; stock −27% in a day. Cause: Autodesk changed its partner incentive model and took the margin. See the warning below. - **Beijer Ref** — CEO stepped down in August 2026 immediately after EQT completed its exit; the CFO is acting CEO with no named successor. Succession is a first-order variable in this model, not a governance footnote. Organic +1% reported, North America −2.9%. - **Topicus** — organic +4–5% is fine; deployment of **€46.3m in H1** is not. The reinvestment rate *is* the engine, and it has stalled. - **Judges Scientific** — profit warning in January 2026 cutting FY26 to 200–250p from ~275p, organic order intake −6% like-for-like with **US orders −23%**, order book down to 15.7 weeks from 18.7, and the acquisition programme effectively halted. The runway argument met a demand shock. - **Roper** — $3.2bn into buybacks, roughly 8% of shares, over three quarters, with the authorisation expanded by another $3bn. Defensible at a de-rated price, but it is a **change in M&A strategy**, which is red-flag category one. - **TerraVest** — Quebec's AMF alleged in June 2026 that the Executive Chairman tipped family and friends ahead of the EnTrans deal, ~C$6.8m of alleged profit. Unresolved as of today. Cheapest multiple in the cohort at 12.4× for a reason. - **DCC** — sold Healthcare and Info Tech, returned £700m, repurchased 13.9% of its shares, and is renaming itself DCC Energy. It has dismantled the model. It is no longer in this universe. - **Storskogen**, **Volati**, **Sdiptech**, **Teqnion** — each interesting, each failing a specific line (turnaround, EBITA −31%, 2.8× leverage, negative organic revenue). ## Three questions the framework cannot answer Everything above is arithmetic. The decision isn't. **1. Is vertical-market software structurally impaired, or is this a sentiment cycle?** This is the whole trade. Constellation's own answer, from Mark Miller on the Q2 call: *"We are not going to give you an AI target, an AI revenue line, or an AI timeline."* If the moat holds, Vitec at 8× and Constellation at 20× are the setup [[Valuing a Compounder]] describes — durable reinvestment mispriced because the market discounts duration badly. If it doesn't, they are melting ice at a fair price. **Nobody has this answer, including me.** Position size is the only honest expression of that. **2. Is +30% organic growth pricing power, or is it the cycle wearing pricing power's clothes?** Amphenol printed +30% organic; NCAB printed +24% with order intake +58% and prices up over 20%. NCAB's CEO says it plainly — AI datacentre investment is straining global PCB supply, and ~20% of Q2 orders were pre-booked for 2027 delivery. Condition three in [[Quality Compounders]] is *pricing power above inflation, independent of the economic cycle.* These two are failing that test on the good side of it, which is the harder direction to notice. **3. Are you being paid for the currency and the rate path?** The Nordic names are a SEK bet you are not hedging, taken while the Riksbank is talking about hiking and the ECB has already moved. FX was a persistent drag on cohort earnings through 2026 and Redeye expects it to persist. ## One warning aimed at your existing book **Addnode is the cautionary tale for anything sitting between a workflow monopoly and its customer.** Its business was reselling Autodesk. Autodesk changed the partner incentive model, took roughly SEK 20m a quarter of margin, and Addnode's EBITA fell 38% with the stock down 27% in a session. That is good news for the ADSK thesis in your book, and a reminder that a workflow monopoly's pricing power is exercised on *whoever is closest*, including its own channel. Before calling any business a digital tollbooth, check whether it owns the road or leases a booth on someone else's. ## Related - [[Quality Compounders]] · [[Acquisition-Driven Compounders MOC]] - [[The Compounder Screener]] · [[Red Flags in Serial Acquirers]] · [[Valuing a Compounder]] - [[Open Questions and Numbers to Reconcile]] · [[Investing Discipline]] · [[Investing Learnings]] --- **Data sources:** company Q2/H1 2026 filings and calls; Redeye *Serial Acquirers Update Q4 2025* (4 Mar 2026); BDO *Nordic M&A Annual Report 2026*; Dealsuite *Nordic M&A Monitor* (Mar 2026); FactSet earnings insight (7 Aug 2026); multpl; FRED DGS10; Riksbank (20 Aug 2026); ECB key rates; Eastspring and S&P DJI on the quality factor; StockAnalysis.com for prices and multiples (several figures carried sign errors or stale dates and were corrected or discarded — treat single-source multiples as indicative, not precise).