JW Pharmaceutical Corp. (JW중외제약) 001060

Deep-value (REFRESH): 001060 — JW Pharmaceutical Corp. (JW중외제약), Korea Exchange

Redo from September 14, 2026, over the reference analysis from August 23, 2026. Reference price … (09/13/2026, tracker). Market cap … (…). Reporting currency KRW = trading currency KRW — no conversion risk in relative metrics.

This redo's regime. The mechanical deep_delta.py triage decided REFRESH on 09/13/2026 for a strictly technical reason: two of the thesis's seven falsifiers were unverifiable, not broken — crestere_venit_yoy needed two consecutive quarters and was missing one, and pe had no value at all. The other five held. In the meantime, the fact that changes everything appeared: the semi-annual report filed August 28, 2026, exactly the filing August's report had nominated as the thesis's test ("that's where you see whether the semi-annual CFO confirms the 56 bn run-rate and whether intangible amortization moves"). With Q2 2026 complete — balance sheet, income statement, cash flow statement — the two falsifiers resolve and, more importantly, the owner earnings base must be recalibrated upward. The delta package's price move (…, from … to …) is context, not a reason.

What I explicitly inherit, without re-verifying: the moat chapter and group structure (JW Holdings → 001060 40.25%; infusion-solution manufacturing sits at the separately listed sister company JW생명과학, not here), the 2022–2025 capital allocation history, the CEO profile and governance file (a suspended criminal conviction in the rebate case, an ongoing tax evasion trial, a fourth term since March 2026, an audit committee under the minimum meeting frequency), the contested FTC fine of …, the mechanics of the sales suspension on 31 products (… = 6.91% of 2025 revenue), and the four-year series of negative accruals. None of these facts has moved in 21 days, and the delta package reports no new filings beyond the semi-annual.

What I fully re-derive: the four quarters through Q2 2026 (balance sheet line by line, margins, cash conversion), the FCF bridge on the TTM window, the owner earnings base, net debt, all five valuation models and the Monte Carlo simulation.

---

Executive summary

The thesis, unchanged in essence, reinforced in the figures. JW Pharmaceutical is a ~… TTM-revenue Korean pharma that completed, over four years, a structural deleveraging (209.5 bn of gross debt and 0.87x D/E in 2022 → 50.1 bn of financial debt and net cash of 52.0 bn in Q2 2026) and trades at 6.0x reported TTM profit, 6.8x tax-normalized TTM profit, 1.50x book value and 4.2x EV/EBITDA. The market hasn't re-rated the transition. What the August 28 semi-annual adds is that the transition hasn't just ended — it's started producing: trailing-four-quarter operating profit is … versus 94.5 bn in FY2025 (…), and operating cash flow is 134.5 bn versus 96.4 bn (…).

The two unverifiable falsifiers, resolved. (1) crestere_venit_yoy required two consecutive quarters. I now have them: Q1 2026 … (199.88 bn versus 185.24 bn, a figure derived as a residual from 2025 annual revenue minus Q2–Q4) and Q2 2026 … (222.02 versus 191.10). The claim under test — "the regulatory shock has been absorbed" — holds, and holds on the quarter fully covered by the sales suspension. (2) pe had no value since yfinance doesn't publish a trailing P/E for this issuer. Calculated from components: market cap 608.8 bn / TTM profit 101.28 bn = 6.01x; on the tracker's EPS (…) = 6.39x; normalized to a 22% tax rate = 6.77x. The claim under test — "the price incorporates almost no growth" — holds with margin. The other five falsifiers stay green: datorie_capitaluri 0.175 (threshold >0.5), dso 48.3 and 52.4 days (threshold >65), marja_bruta 47.42% and 49.26% (threshold <0.42), pret … (threshold >40,000), roe 24.7… (threshold <0.05).

The recalibration that matters: owner earnings from 56 to …. The old base was FY2025's FCFE (55.51 bn), rounded to 56. That exercise contained a one-off payment of ~… from the 2021–2024 tax recalculation. On the TTM window ending Q2 2026, the poisoned quarter has left the series, and gross FCFE is 95.3 bn. That figure is itself flattered, in the opposite direction: the TTM effective tax rate is 12.2%, half of normal, due to provision reversals in Q3 and Q4 2025. Normalized to 22% and loaded with the working-capital increase the TTM window happened not to pay, the declared base is … — 24% below gross TTM FCFE, 29% above the old base, and corresponding to an FCFE yield of 11.8% versus the …% simple FCF yield the data pack independently publishes. The full bridge is in the valuation chapter.

Estimated value. With OE …, growth 6% for five years and 3% for years 6–10, a discount rate of … (unchanged — the 1.8pp governance penalty over CAPM remains fully applied) and terminal growth of 1.5%, intrinsic value is …, i.e. a margin of safety of …. Triangulated across five models, the range is … … …, with a median of …. Monte Carlo over 20,000 scenarios, with the same central assumptions: P10 …, P50 …, P90 …, probability of undervaluation …%. The single most important number remains, as in August, the Earnings Power Value one: with no growth ever, the business is worth …, i.e. … above today's price — in August the same method gave …. The sign flip at EPV is the direct consequence of rebasing, and it's where the thesis is most vulnerable to an error of mine.

Verdict: BUY, average position — sized on governance, not percentiles. The business is visibly better than the price, earnings quality remained HIGH across all five O'Glove axes, and the semi-annual validated both signals the August report had left to watch: receivables fell instead of rising (117.8 bn at 06/30/2026, DSO 48.3 days, versus the 128 bn / 65-day alarm threshold), and intangible amortization rose from 0.50 to 0.73 bn per quarter. What hasn't moved at all is key-person and control risk: the same CEO with a criminal conviction and an ongoing tax trial, the same audit committee that doesn't meet quarterly, the same unquantified intra-group transactions. At 6x profit you're paid to bear that; the simulation's …% probability varies assumptions, it doesn't vary whether the reported figures are the real ones.

---