Old Republic International Corporation ORI

ORI — Old Republic International Corporation — Deep-value report (REFRESH)

Generated: 2026-09-04 · Price: … (close 03.09.2026, yfinance) · Market cap: …M · Shares issued and outstanding: 241,432,922 (30.06.2026, 10-Q p.3) · Exchange: NYSE — no BSE collision, query_tracker indicatori ORI confirms exchange=NYSE · Reporting currency USD / price USD, no ADR, no conversion.

Regime: REFRESH over the thesis from 2026-07-27 (delta pack delta-ORI-20260903.md, REFRESH triage on 1 new filing). The source triggering the revisit: 10-Q for Q2 2026, filed 31.07.2026 (accession 0000074260-26-000092). The local file set contained no 2026 filing besides the 10-K — edgar_10k_downloader.py only downloads 10-Qs for the fiscal year covered, so I manually downloaded from EDGAR and generated text+index with filing_text.ruleaza('ORI'): SEC-Filings\ORI\ORI_10Q_2026-07-31.txt (1,561 lines) and ORI_10Q_2026-05-01.txt (1,354 lines).

What I inherit unchanged from the 27.07.2026 report (verified the new filing doesn't contradict it): the two-segment business structure after the RFIG sale to Arch Capital (31.05.2024, …M proceeds — confirmed in the FY2024 cash flow, "Proceeds from sale of subsidiary 136.6"); the moat description and the 19-autonomous-operating-companies model; the 84-year uninterrupted dividend history and 44 consecutive years of increases; the Specialty line mix for 2023-2025; the investment portfolio composition and the 2025 rating-mix shift; the management team (Smiddy CEO since 2019, Sodaro CFO since 2021); the absence of LDTI, the disclosed DAC policy, the absence of goodwill impairment, the absence of material non-claim litigation; reserve development for 2023-2025 and discount accretion on workers' comp.

What I fully re-derive, because the new filing moved the facts: recurring earning power and the FCFE bridge, earnings quality (O'Glove), the full valuation (all five models + Monte Carlo), the capital allocation profile, and the verdict.

Primary sources: 10-Q Q2 2026 (31.07.2026) and 10-Q Q1 2026 (01.05.2026), 10-K FY2025 (26.02.2026), FY2024 (27.02.2025), FY2023 (28.02.2024), 10-Q Q2 2025 (01.08.2025) — all in SEC-Filings\ORI\; deterministic data pack rapoarte\deep\data-pack-ORI-20260904.md; research brief rapoarte\deep\research-ORI-20260904.md; simulation rapoarte\deep\mc-ORI-20260904.json.

---

Executive summary

The thesis has flipped from "cheap on P/E, expensive on P/B" to "expensive on any recurring-earnings metric." Not because of price — the price is practically unchanged from the July analysis (… → … …) — but because the Q2 2026 10-Q shows the engine that held margins up has stopped.

The central finding, quantified. Operating income (net income excluding net investment gains, the metric management itself uses) fell in the first half of 2026 to …M from …M (…), and per diluted share to … from … (…) — 10-Q Q2 2026, p.23. I built the pretax bridge from …M to …M line by line (see the "What changed" chapter): of the …M lost, …M comes exclusively from the disappearance of favorable reserve releases at Specialty (-3.1pp from the loss ratio in H1 2025 → -0.6pp in H1 2026), and in the second quarter consolidated development fell to 0.1 favorable points versus 2.1 points last year. The reserve cushion that the July report described as "halved since 2023" hasn't halved further — it's evaporated.

Second finding: the multiple reported by the tracker is an artifact. The ttm P/E of 9.23x and ROE of 18.6% from the data pack are calculated on GAAP profit, which in the first half of 2026 contains …M of after-tax investment gains (…M of which is pure mark-to-market on the equity portfolio, with no cash flow — 10-Q p.4 and the CFO reconciliation, p.7). On TTM operating earnings of …/share, ORI trades at 14.3x, not 9.2x — i.e. above its own 5-year average (~10.7x) and above the comparison group (~13.5x) cited in July. The mechanical falsifier "P/E 9.15 < 13.5" only passed because it reads GAAP EPS; economically, it's broken.

Key figures at 30.06.2026 / 03.09.2026. Consolidated combined ratio H1 96.0% (vs 93.7%); Specialty 95.2% (vs 90.2%) — at the upper edge of management's declared target band, 90…; Title 97.4% (vs 100.4%), the one real piece of good news. Book value/share … (… from end of 2025). Annualized operating ROE 11.5% (Q1) and 12.1% (Q2), versus ~13.8% in 2025. P/B 1.68x, but 2.06x if you strip the accumulated unrealized gain on the equity portfolio out of equity (…M gross, ~…M after tax). Total debt …M versus …M at 31.12.2025 — a …M 5.700% senior notes issuance in May 2026, prefunding for the …M maturity in August 2026 (real net increase ~…M, not …M).

Estimated value. Five triangulated models give a margin of safety of … to …, median at … — central intrinsic value ≈ …/share, reasonable band …-… The Monte Carlo simulation (20,000 scenarios, assumptions declared below) gives a median of … and an undervaluation probability of …% — practically a coin flip. The range is wide (P10 …, P90 …) because the owner-earnings base is itself in motion.

Verdict: UNINTERESTING / AVOID AT THIS PRICE — downgraded from INTERESTING (27.07.2026). The company remains high quality (conservative balance sheet, disciplined distribution, long-tenured internal management), but at … you're paying 14.3x declining operating earnings and 1.68x a book value 22% of which is a US equity portfolio marked to market. It's not a short thesis and not a broken company — it's a sound company at a price that no longer leaves room for error, exactly when the number of possible errors has grown. Prior GBL score from tracker: 68.6% — INTERESTING (evaluated 27.07.2026); the divergence from my verdict is explained in the final chapter and comes entirely from the scorecard reading GAAP EPS and ROE.

---