CRN — Cairn Homes plc

Deep-value report: CRN — Cairn Homes plc (REFRESH)

Date: 2026-09-03 · Price: … (LSE, CRN.L, close 2026-09-02) ≈ … (GBP/EUR 1.1631) · Market cap: ≈…M / …M · EV: …M (H1 2026 net debt …M) · Exchange: London Stock Exchange (secondary listing) + Euronext Dublin (primary, C5H.IR) · Sector: Consumer Cyclical / Residential Construction (Irish homebuilder) · ISIN IE00BWY4ZF18

REFRESH regime. The reference analysis is from 2026-07-26 (verdict: INTERESTING, 68.3% GBL). The mechanical deep_delta.py triage called for a re-run for 2 unverifiable falsifiers (operating margin, revenue growth) — both needed a new filing to be testable. That filing has since appeared: the H1 2026 interim results, published 2026-09-02, the company’s first complete financial statement since the reference thesis.

What I explicitly carry over from the 07/26/2026 report, since nothing material moved: the company history and business structure (founded 2014, IPO 2015, integrated homebuilder model), the moat analysis (a listed duopoly with Glenveagh in a market of 434 builders, LDA partnerships, the lack of a real switching cost), the 2019-2025 capital-allocation track record, the governance structure and auditor (EY), Irish public-policy risk as a theme, and the resolution of the symbol collision (CRN = Cairn Homes, CRN.L; the company does NOT file with the SEC — there’s no CIK, so the EDGAR step produces no filing, the primary sources are the company’s RNS releases/results

• yfinance).

What I fully re-derive, since the new filing moved the facts: the FCF bridge (now on real H1 2026 figures, not an assumed conversion), the owner-earnings base, all five valuation models, the Monte Carlo simulation, the quality-of-earnings chapter (O’Glove) and the CEO profile chapter (Thorndike) — the last two were entirely missing from the reference report and are written here from scratch.

The price move (… → …, … in 38 days) is context, not thesis. I build nothing on it; I use it only where it matters arithmetically — the margin of safety.

⚠️ DIFFERENT currencies. Cairn reports in EUR, the stock is quoted in London in GBp (pence). All intrinsic values below are calculated in EUR and converted to GBp using fx = 0.011631 (EUR per 1 GBp, i.e. GBP/EUR 1.1631 at 2026-09-03). I re-checked this time the yfinance indicators too: trailingPE …x, priceToBook 2.02x and bookValue …/share are correct and consistent (book value expressed in GBP, not EUR) — I don’t repeat the P/B error warning from the previous report, which concerned only enterpriseValue. Correctly recalculated EV manually: …M, not the yfinance figure.

Executive summary

The H1 2026 results (published 2026-09-02) did exactly what was needed to settle the thesis: they turned two unverifiable statements into measured facts. The outcome is asymmetric — one of them was confirmed spectacularly, the other broke.

The volume falsifier held, and with margin. H1 2026 revenue was …M, … against …M in H1 2025, on 1,139 units delivered (… versus 708). Operating profit rose …, to …M; pre-tax profit …, to …M; basic EPS …, to 9.3 cents. FY2026 guidance was raised for the second time in six months: revenue ~… (from the …-1.08bn range) and operating profit ~…M (from …-185M), with guided ROE raised from 16.5% to 17.0%. The order book stands at 5,020 units / …, over 900 more units than last year, and forward sales (3,881 units / …) cover WIP investment by 3.0x. The … output ramp-up for 2026-2027 is no longer guidance — it’s delivery in progress.

The margin falsifier broke. The reference thesis said operating margin was “on a steady climb” (16.7% → 17.0% → 17.4% → 17.9% in 2022-2025). FY2026 guidance implies 185.0/1,080 = 17.13% — the first operating-margin decline in five years, −72bp against FY2025. The cause is visible in H1: gross margin fell to 21.3%, from 22.1% (−80bp), while execution costs (opex) continued to dilute normally. The engine is a price/cost scissors — net ASP grew only … (… → …) while construction-cost inflation is guided at 2.5% for 2026 (versus ~1% in 2025), on a mix pushed toward cheaper products (duplexes 28% and apartments 9% of sales, at … and … against … for houses). The honest nuance: 21.3% is at the low end of a historical 21.7… band, not below it — the margin is flattening, not collapsing — but the “steadily climbing” claim is dead and can no longer be used as a value driver.

What moved the valuation most isn’t either of those — it’s cash flow. The reference thesis used owner earnings of …M, derived from an assumption — 2026E net profit …M × an assumed “compressed” conversion of 55%, because WIP was expected to keep swallowing cash. H1 2026 shows the assumption was conservative for the wrong reason: net WIP investment fell from …M to …M (…), and operating cash flow flipped from −…M to +…M, a …M turnaround in a year. CFO Richard Ball said explicitly on the call that he no longer sees “the material net land investment that we would have seen in the last 24 months.” My re-derived base is FCFE 2026E = …M — and the check that matters: the average realized FCFE over the last four years (2022-2025) is …M, so the figure isn’t a projection dressed up as fact, but also the average of what the company has actually delivered.

Estimated value (range, five triangulated models): from … (conservative DCF, WACC 10%/g1 7%/gt 1.5%) through … (historical multiple, 10.2x on 2026E EPS) and … (EPV Greenwald at WACC 8.7%) up to … (central DCF) and … (bull DCF). The median of the five sits at … below price; Monte Carlo across 20,000 scenarios gives a median of … and only …% probability the stock is undervalued. Broker consensus (4 analysts, MarketScreener) is at … = …, and the most bullish published target (Berenberg, …) is … — my central case of … is above everything the market publishes, which is a signal of caution, not of discovery.

Verdict: INTERESTING / WATCH — and less attractive than on July 26. The company delivered better than expected on volume and cash, but the price rose faster than the value: the margin of safety in the central case narrowed from … to …, and the growth-free floor (EPV + historical multiple, 209-…) is now 7… BELOW price, against “practically level” in July. Today you’re buying real quality (17% ROE, 18.6% D/GAV, no maturities before June 2029) at a price that already prices in guidance execution. Convergence with the prior GBL score (68.3% — INTERESTING): strong, with the caveat that the argument supporting the upper end of the range (margin expansion) has disappeared.