Recovery already priced after the +122% run · −24% margin of safety
Market in the top decile of the MC band — P(intrinsic<market) = 94.5%Intrinsic value ₩22,238/share (after a 10% governance discount; pre-discount ₩24,709) vs market ₩29,250. The Monte Carlo median is ₩20,903 and only the p95 tail (₩29,519) reaches today's price — 945 of 1,000 correlated draws land below it, so the recovery the stock already ran on is fully discounted, not a margin of safety.
What it sells, where it sells
Operating segments
Single-segment in the filings: ~94% is one cyclical display-FPCB line concentrated in Apple and Samsung Display, so the entire valuation lives in one cell — the through-cycle operating margin on that line — and the auto/robotics/IT-OLED adjacencies are forward optionality, not yet cash flow.
Country mix (revenue-weighted CRP input)
Weighting is production-location: Vietnam books the bulk of the Samsung-Display / Apple FPCB volume while the US is the Apple end-demand — so this is a KRW-reported DCF of a USD-priced exporter, and the blended ~2.7% CRP add-on over the US mature ERP carries the Vietnam/China manufacturing risk.
Background — five things to know before the case
Context the share price doesn't carry on its face. Skim this once and the bull / bear bullets below stop reading as inside baseball.
- #1 Korean and ~30%-global FPCB maker — but a single-customer commodity supplier. ~94% export, concentrated in Apple (OLED-RFPCB since 2017) and Samsung Display. Oligopoly share among sellers does not confer pricing power when you sell to a monopsonist buyer; the surplus accrues to Apple, not to BH.
- The bear case is entirely in the margin, not the top line. Revenue plateaued ₩1.6–1.8T 2022–2025 after a 5× run off the 2016 trough. Operating margin compressed from a ~7.8% 2022 peak through 5.3%/5.0% to a ~3% FY2025 trough (OP ₩54bn), recovering to ~4.9% TTM — net income fell ₩144bn → ₩31bn over the same span.
- Active dilution is the flagged warning sign. Share count +5.3% YoY; a ₩27bn Series-4 zero-coupon convertible (100% convertible at ₩21,600, in-the-money at ₩29,250, ~1.25M / ~3.7% latent shares, 2030 maturity) sits against a buyback-and-cancellation (752,841 shares retired Sep 2025). Net: dilution currently outpaces cancellation.
- The stock has already run +122% in 12 months (+40% in a week). Daishin models FY2026 OP ₩126.8bn (+135% YoY); the 11-analyst average PT (~₩28k) was set before the latest run, so the market is now ahead of consensus and pricing the recovery the bulls describe.
- Founder is a ~21.5% minority-of-record, not a controlling holder. The rest is institutional/foreign (Morgan Stanley ~5.9%, NPS ~4.2%) + 56% public. This is a cyclical opco DCF with a dilution overlay — not a holdco NAV-trap analysis. The 24.89% DK Tech cross-holding is carried at ₩66bn (≈35% haircut + ~20% look-through CGT), not the ₩123.7bn screen value.
Where we diverge from the sector
Each row asks: if we'd used the sector median for this one assumption instead of our override, how much would the share price change? Negative = our override is more conservative (worth less); positive = more aggressive (worth more). Skim down the rightmost column to see which bets are doing the work.
The story & the five claims
The 10-year story this DCF is built on, plus the five anchor claims that translate the story into model inputs.
The 10-year story
BH is the #1 Korean and roughly 30%-of-the-world flexible-printed-circuit maker — a deeply cyclical, capital-heavy contract supplier whose entire value is the through-cycle cash flow of one display-FPCB line concentrated in Apple and Samsung Display. The top line is not the problem: revenue plateaued in a ₩1.6–1.8T band for four years after a 5× run off the 2016 trough. The whole debate is the margin, which fell from a ~7.8% peak in 2022 to a ~3% trough in 2025 and is now recovering toward ~4.9%. The base case is a trough-to-mid-cycle normalization, not a re-rating: revenue grows about 8.5% a year for five years off the recovering base as foldable-iPhone, robotics and auto-charging stack on the iPhone FPCB business, then fades to a 3% terminal; operating margin normalizes to a year-10 6.5%, inside the band BH actually printed in 2021–22 but explicitly below the 7.8% peak. Terminal returns are pinned at the cost of capital — an honest steady state for a single-customer commodity supplier, not a moat rent. After a 10% governance discount for forward dilution, the intrinsic value is ₩22,238 against a ₩29,250 price: a stock that has already run +122% in twelve months has fully priced the recovery this model assumes.
Two debates worth pressure-testing
- Base operating income re-based from TTM ₩90bn to a normalized ₩72bn. The TTM print removed a real −₩25.3bn Q1'25 loss quarter and added a recovery quarter — the highest 4-quarter sum on the calendar, 67% above FY2025 actual. Sell-side and the screen anchor on the higher number; we anchor on the through-cycle blend.
- Terminal ROC pinned to WACC, not to a 10% moat rent. The flow triple (6.5% × 1.8 × (1−0.24)) derives ~8.9% ≈ the 8.54% WACC. A company whose ROIC has been compressing does not earn above its cost of capital by assumption — pinning it would manufacture excess returns the history denies.
- The recovery is counted once, not twice. Sell-side narratives double-count it — once in an elevated base, again in a 3%→6.5% margin ramp. We pick one: a normalized base plus the ramp, which is what re-basing enforces.
- β set to the 5Y regression 0.91, not the industry bucket 1.93. The Damodaran Electronics (General) β_u 1.57 re-levers to 1.93 — a 112% divergence from the regression. The bucket is a volatile small-cap EMS aggregate; BH's large swings are idiosyncratic Apple/robot narrative, not market-correlated, so a moderate market β is correct (caveat: the lived 200% swing argues the realized risk is higher).
- DK Tech carried at ₩66bn, not the ₩123.7bn screen value. A speculative humanoid-robot stake takes a ~35% volatility/illiquidity haircut net of ~20% look-through CGT; and the in-the-money CB's ~1.25M shares are added to the diluted count now rather than treated as contingent.
- The trough is the new normal. If Apple presses price on a 94%-concentrated book and the new vectors slip, mid-cycle margin stays ~3-4.25% rather than recovering to 6.5%. The bear scenario lands ₩10,676 — a −64% drawdown from today's price.
- The whole upside rests on three pre-commercial businesses. Foldable, humanoid-robotics FPCB and BH EVS auto-charging “barely have revenue” — a venture bet wearing a DCF's clothing. Strip them out and the core display-FPCB line does not clear the price.
- Customer concentration is a single point of failure. One Apple or Samsung Display sourcing decision vaporizes the thesis — a tail a 0.91 beta does not capture. The stock has moved 40% in a week and swung ₩11,910→₩38,500 over the window.
- Dilution is live, not “forward risk.” Share count +5.3% YoY plus the ₩27bn convertible; net dilution currently outpaces the buyback-and-cancellation, so per-share value leaks even as the business recovers.
- FX / cycle-phase mismatch. A KRW-discounted DCF of a USD-invoiced exporter, and an 8.5% fade lands the terminal value mid-to-late cycle — so the perpetuity risks capitalizing a non-normalized margin regardless of the base chosen.
Risks to thesis (tail, not bear case)
The +122% 12M run means the bull margin recovery is in the price. A DCF that re-derives ₩29,250 by assuming that recovery offers no margin of safety — the central reason this is a SELL/PASS, not a Buy.
94% to two monopsonist buyers means the buyer sets the margin. If mid-cycle is structurally 3-4% not 6.5%, the bear ₩10,676 (−64%) is the path. This is the one cell the whole valuation turns on.
Terminal value is 94.8% of EV — the explicit decade is FCFF-negative through Y5 as the foldable/robot/IT-OLED lines fund. A cyclical supplier ramping at the bottom plows cash into working capital and capacity exactly when the model banks the recovery.
+5.3% share count YoY and a ₩27bn in-the-money convertible; net dilution outpaces cancellation. The 10% governance haircut prices the forward propensity; the existing CB shares are already in the count.
KRW-discounted flows against a USD-invoiced Apple book; the won risk-free is cyclically elevated by the oil shock. A stress to 3.0-3.3% rf and forward-FX reconciliation move the answer.
The bull case — foldable + robotics ignite a second super-cycle and level margins to 7.8%+ — is real but unproven. It carries ₩41,158 (+41%) at 15% weight; if it converts, the SELL is wrong, which is why this is a context-check, not a conviction short.
10-year forecast
Revenue ₩1.99T → ₩3.55T over 10y (~8.5% Y1-5 off the recovering base, fading to 3.0% terminal). Operating margin lifts from ~4.4% Y1 to 6.5% by Y5 and holds — but FCFF is negative through Y5 as the capital-heavy lines fund, so 94.8% of EV is the terminal block.
Monte Carlo distribution
The median Monte Carlo outcome (₩20,903) sits 29% below today's ₩29,250, and only the p95 tail (₩29,519) reaches the price — 945 of 1,000 correlated draws land below it, so P(intrinsic < market) = 94.5%.
Mean ₩20670.32 ± ₩5291.74/sh, 1000 iterations (0 failed). P(intrinsic < market ₩29250.00) = 94.5%.
Cost of capital build
| Risk-free rate | 3.78% |
| Mature-market ERP | 3.97% |
| Levered β | 0.91 |
| Weighted CRP | 2.66% |
| Cost of equity | 10.05% |
| Pre-tax cost of debt (synth A1/A+) | 4.48% |
| D / V | ~23% |
| WACC | 8.54% |
Full year-by-year DCF
| Year | Revenue | Op mgn | EBIT | EBIT(1−t) | Reinvest | FCFF | PV |
|---|---|---|---|---|---|---|---|
| 1 | ₩1.99T | 4.44% | ₩88.33B | ₩66.25B | ₩111.22B | ₩-44.97B | ₩-41.43B |
| 2 | ₩2.16T | 4.96% | ₩106.92B | ₩80.19B | ₩120.67B | ₩-40.48B | ₩-34.36B |
| 3 | ₩2.34T | 5.47% | ₩128.04B | ₩96.03B | ₩130.93B | ₩-34.90B | ₩-27.29B |
| 4 | ₩2.54T | 5.99% | ₩151.96B | ₩113.97B | ₩142.05B | ₩-28.08B | ₩-20.23B |
| 5 | ₩2.75T | 6.50% | ₩179.03B | ₩134.28B | ₩154.13B | ₩-19.85B | ₩-13.18B |
| 6 | ₩2.96T | 6.50% | ₩192.28B | ₩144.60B | ₩113.24B | ₩31.36B | ₩19.15B |
| 7 | ₩3.14T | 6.50% | ₩204.40B | ₩154.12B | ₩103.54B | ₩50.58B | ₩28.34B |
| 8 | ₩3.31T | 6.50% | ₩215.03B | ₩162.56B | ₩90.84B | ₩71.72B | ₩36.82B |
| 9 | ₩3.44T | 6.50% | ₩223.84B | ₩169.67B | ₩75.35B | ₩94.32B | ₩44.29B |
| 10 | ₩3.55T | 6.50% | ₩230.56B | ₩175.22B | ₩57.40B | ₩117.83B | ₩50.51B |
Methodology & flags
Damodaran FCFF DCF, 10y explicit + perpetuity, in KRW. Risk-free
3.78% (local-currency government bond). Synthetic credit A1/A+. CRP
2.66% from revenue-weighted country mix × Damodaran 2026 CRPs.
Terminal ROIC faded to 8.54%; 10% governance haircut applied
post-DCF. Monte Carlo: 1000 iterations. Engine v1.0.0 · result:
/Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/bh/output/2026-06-02-result.json
- R&D cap: OFF · Lease cap: OFF · Failure: OFF · ESO: OFF
- Governance haircut: 10% applied post-DCF (₩24709.19 > ₩22238.27)
- Sensitivity tornado: not run