← back to picks
HQ Seoul · South Korea Listing KRX:044820 Reporting KRW Structure holding company · sum-of-the-parts Damodaran FCFF · 1000-iter MC FCFF · Dark v3

Realizable cash to a minority ≈ ₩13,841/sh vs ₩18,120 market · −24% · NAV ceiling ₩30,766 is a vault you don’t hold the key to

SELL — market price exceeds the cash a minority can realize; the NAV backing is trapped

Cosmax BTI is the listed holding company of Korea’s Cosmax group, and it is a dual-lens valuation. The realizable number — the cash a minority actually receives (a ~2% dividend, growing ~3.5%, that is not even covered by FCF, plus a small catalyst option) — is worth ≈₩13,841/share, 24% below the ₩18,120 market price (a cash-to-minority dividend look-through). The NAV ceiling — what the assets are worth to the controller — is ₩30,766 (the DCF NAV number, after modelling capital laziness in the flows, marking the trapped 27% Cosmax Inc stake control-net to ₩435B, and a residual 15% governance haircut). The ₩16,925/sh gap between the two lenses is the trap: Korean controlled holdcos trade below NAV for decades, and a minority can be right on NAV and collect nothing. On the realizable lens this is a Sell pending a named catalyst (buyback-and-cancellation, holdco/opco merger, or activist/NPS ≥5% filing), not a clean buy. This rebuilds v1, which mis-stated a +96% “buy” by carrying the stake gross and bolting a flat 45% discount on the end.

p5 ₩27.4k p25 ₩32.8k p50 ₩37.0k p75 ₩41.2k p95 ₩46.7k ◂ MARKET ₩18.1k (off-scale left) NAV ₩30.8k
StructureListed holdco · sum-of-the-parts / NAVCrown jewel27.24% of Cosmax Inc (KRX:192820) · ₩435B at marketOpcoNBT health food + packaging · ₩338B EVβ / MC σ1.78 re-levered (β_u 0.56) · ±₩5.9k/sh (1000 runs)GovernanceControlling family ~53.5% · 15% haircutCaveatNAV ceiling · no catalyst · value-trap risk
Intrinsic / share
₩13,841.00
cash-to-minority · NAV ceiling ₩30,766 (post 15% gov)
Market / share
₩18,120.00
late-May 2026 · KRX:044820
Margin of safety
-23.6%
vs intrinsic
Enterprise value
₩337.57B
71.6% terminal
Cost of equity / debt
11.60% / 4.05%
β 1.78 · CRP 0.70%
Terminal ROIC / g
6.00% / 2.50%
spread ~5bp (ROIC 6.00% vs WACC 5.95%)

What it sells, where it sells

Operating segments

₩645.2B FY25 consol. revenue
Cosmax NBT — health functional foodSeparately listed (KOSDAQ 222040). Soft domestic Korean HFF demand offset by NBT supplement export tailwind (US / SE-Asia) riding the K-wellness wave~70%
Cosmax Neo / CM Tech / logisticsCosmetic packaging, raw materials, logistics & equipment, overseas marketing — commodity, low-margin, capital-heavy~30%

This donut is only the directly-consolidated operating businesses — the ₩338B enterprise value the FCFF DCF actually prices. The crown jewel (the 27.24% Cosmax Inc associate) is equity-method, so its ₩31B of income sits in non-operating investment income and its ₩435B market value is carried separately as a cross-holding, not in this revenue line.

Country mix (revenue-weighted CRP input)

🇰🇷South Korea75%
🇺🇸United States12%
🇨🇳China8%
🇮🇩Indonesia5%

This is the geography of the CONSOLIDATED OPCO only — Korea-dominant, with the ~25% foreign weight being NBT supplement export demand (US / China / Indonesia). It deliberately does not reflect the China/US-heavy Cosmax Inc associate, which is a separate cross-holding marked to market; mixing the two would double-count the group's foreign exposure.

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.

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.

Assumption
Our input
Sector median
Value impact
01Opco revenue growth10-year consol. revenue CAGR vs Food Processing median
~3% → 2.5%NBT export tailwind partly offsets soft domestic HFF and commodity packaging — not a growth story
~4%Food Processing, global cross-sector median
small
02Opco operating marginYear-10 target vs Food Processing median EBIT margin
5.0%Modest lift from 4.2% on operating leverage / export mix — kept well below the sector median
~8.5%Food Processing global median operating margin
conservative
03Cross-holding (the crown jewel)27.24% Cosmax Inc stake — carried at market, not re-DCF'd
₩435B₩680B gross (3.09M sh × ~₩220k) × 0.80 look-through CGT × 0.80 control-realization — a trapped 27% minority stake, not gross NAV
n/aNo sector default — this is the load-bearing line
+₩435B
04Beta (anti-suppression)Re-levered Food Processing β_u vs the suppressed regression β
β_lev 1.78β_u 0.5602 (Damodaran Food Processing) re-levered at the real consol. D/E ~2.97 — the 0.65–0.78 regression β is discarded as a microcap-illiquidity artifact
~0.68Raw regression β — suppressed by 3.5M-share float and stale pricing
conservative
05Governance / holdco haircutStacked Korean holdco-NAV + controlling-family discount
15%Residual leakage/dilution only — governance is carried in the flows (terminal ROC ≈ WACC, zero excess return) and the control-net stake mark; below the navien/daewoong 0.25 tunnelling ceiling
~0–30%Damodaran typical band for controlled entities
−₩5.4k/sh
Net effect of the structure
The number is dominated by two lines, not the opco DCF: the ₩435B cross-holding that lifts statutory NAV to ₩36,195/share, and the 15% residual governance haircut that cuts it to ₩30,766 — with capital laziness already modelled in the flows (terminal ROC ≈ WACC). Everything in the FCFF model (growth, margin, β) moves the ₩338B opco EV — a rounding error next to the stake. Read the result as a NAV ceiling; the cash a minority actually realizes is ≈₩13,841/sh (cash-to-minority lens), 24% below market.
₩30.8k
Our override Sector median Adds value Subtracts value

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

Cosmax BTI is not really an operating company you value with a single DCF — it is the listed holding company of Korea's Cosmax group, and its worth is a sum of two very different parts. The first part is the small, heavily-indebted set of businesses it actually consolidates: health-functional-food maker Cosmax NBT, cosmetic packaging, raw materials and logistics. Together those did ₩645B of revenue last year at a thin 4.2% operating margin, with earnings that have shrunk about 9% a year for five years; the FCFF model values that whole pool at only ₩338B of enterprise value. The second part — and the reason anyone looks at this stock — is a 27.24% stake in the separately-listed Cosmax Inc, the ₩2.4 trillion-revenue ODM that is currently firing on all cylinders. That stake is worth about ₩680B gross at today's market price; net of ~20% look-through capital-gains tax and a 20% control-realization haircut — BTI holds only a 27% minority of it, and cannot force it to pay or sell — we carry it at ₩435B. Add the stake to the opco, subtract the group's ₩407B of net debt, and statutory net-asset value comes to roughly ₩36,195 a share. Then we apply only a 15% residual haircut — because the governance load already lives in the flows (terminal ROC set to its cost of capital, so reinvested capital earns zero excess return) and in the control-net stake mark, leaving the end-discount to price residual leakage and dilution alone — to land at ₩30,766. Against a ₩18,120 price that NAV ceiling still looks like a +70% gap, but it is a ceiling, not a target: the cash a minority actually receives is worth ≈₩13,841/sh — 24% below the price — the discount has every reason to stay open, and without a buyback or a NAV-closing event a minority holder can wait a very long time to collect it.

Two debates worth pressure-testing

Is the NAV gap a real margin of safety?
Our view: No — not on the cash a minority receives. The NAV ceiling (₩30,766) is what the assets are worth to the controller; the cash a minority can realize is ≈₩13,841 (a thin, FCF-uncovered dividend + a small catalyst option), 24% below the ₩18,120 price. Korean controlled holdcos trade below NAV for decades — the gap is real on paper and uninvestable in cash.
Why a 15% residual discount, not v1’s 45%?
Our view: Because governance now lives in the flows and the asset mark, per the controlled-company methodology: terminal ROC ≈ WACC (zero excess return on reinvestment) and the 27% Cosmax Inc stake carried control-net (₩680B × 0.80 CGT × 0.80 control-realization = ₩435B). The flat 45% was the whole thesis hiding in one knob and double-counting CGT. The residual 15% prices only leakage/dilution — below the navien/daewoong 0.25 tunnelling ceiling.
What flips it to a buy?
Our view: A named catalyst that routes trapped NAV to outside holders — a declared buyback-and-cancellation, a holdco/opco merger, or an activist/NPS ≥5% disclosure. Until then the ₩30,766 ceiling is academic.
CLAIM 01The opco is a low-growth, low-margin, levered side-show.rev growth ~3% → 2.5% · op margin 4.2% → 5.0% · EV ₩338BFY25 consol. revenue ₩645B at 4.2% margin, earnings down ~9%/yr over 5y, net interest cover ~1.0x. The FCFF model prices this pool — and it is a rounding error next to the stake.
CLAIM 02The crown jewel is a ₩435B cross-holding, not operating cash flow.cross_holdings: ₩435,000M · 27.24% of Cosmax Inc, control-net₩680B gross × 0.80 look-through CGT × 0.80 control-realization. Equity-method and a minority stake BTI cannot force to pay or sell — so it is marked to what a BTI minority realizes, not gross NAV.
CLAIM 03The equity bridge nets ₩407B of net debt and ₩20B minorities.EV ₩337.6B − debt ₩513.9B − minority ₩20.2B + cash ₩107.2B + cross ₩435B = ₩345.7BPre-governance equity ₩345.7B = ₩36,195/share. Consolidated leverage is real, but the ₩435B stake backstops the debt — going-concern failure is unlikely (failure prob set 0%).
CLAIM 04Beta is re-levered from the sector, not the suppressed regression.β_u 0.5602 → β_lev 1.78 at D/E ~2.97 · WACC 5.95%The 0.65–0.78 regression β is a microcap-illiquidity artifact (3.5M-share float, stale pricing). Damodaran global Food Processing β_u re-levered at the real consol. D/E gives a defensible 1.78.
CLAIM 05A 15% residual haircut — NAV is a ceiling, not a target.governance_discount: 0.15 · ₩36,195 → ₩30,766/shareResidual leakage/dilution only: governance is already carried in the flows (ROC ≈ WACC) and the control-net stake mark. Below the navien/daewoong 0.25 tunnelling ceiling. The realizable cash-to-minority is ≈₩13,841/sh — the number that governs the verdict.
Where we diverge from sell-side
  • We treat the number as a NAV ceiling, not a price target. Sell-side that looks at this name anchors on the statutory NAV gap; we model governance in the flows, mark the stake control-net, apply only a 15% residual haircut — and still flag that the realizable cash-to-minority (≈₩13,841) sits below market, so the NAV gap may never close without a catalyst.
  • Crown jewel carried at market, not re-DCF'd. We resist the temptation to value Cosmax Inc higher — marking the stake to a conservative ~₩220k keeps the thesis honest and the debate on the haircut.
  • β set to 1.78 by sector anti-suppression, not the ~0.68 regression. A 0.68 equity β at D/E ~3x implies β_u ≈ 0.21, below any food company — clearly an illiquidity artifact, so we re-lever the sector unlevered β instead.
  • Governance modelled in the flows, not a flat end-discount. Capital laziness enters as terminal ROC ≈ WACC (zero excess return) and the 27% stake is marked control-net; the residual end-discount is just 15%, well inside the 0–30% band. This corrects v1, which crammed the whole thesis into a flat 45% knob.
  • Opco WACC is flattered and we say so. Consolidated book debt loads the debt weight to ~0.75, pulling WACC to ~6%; much of that debt actually carries the Cosmax Inc stake, so the opco EV is an upper-ish estimate. It barely matters — the stake and net debt dwarf the opco DCF.
Two-sided case — bear anchors
  • The discount never closes — the value-trap case. This is the central bear case, and the one the 0% probability-below-market cannot capture. Korean controlled holdcos trade 50–70% below NAV for decades. No buyback, no NAV-closing event, management indifferent: a minority holder can be right on value and still collect nothing for years.
  • The controlling family routes cash for its own ends. The holdco is a succession vehicle; the May 2026 control change ran through private shells, and there is a history of dilutive discounted issuance to minorities. The 15% haircut prices this; a worse-than-modelled related-party drain would justify even more.
  • The crown jewel re-rates down. ₩435B of the value is one marked-to-market stake. Cosmax Inc trades on a China re-acceleration and a US breakeven inflection; a China stumble or a strong-KRW translation hit would cut the stake — and the Monte Carlo already varies it across the ₩155k–₩287k 52-week range.
  • Opco leverage bites in a downturn. ₩513.9B debt, ~1.0x interest cover, a dividend paid above opco FCF. If the associate pickup falters, the consolidated balance sheet is fragile and the ~Caa/CCC synthetic rating is not cosmetic.
  • No catalyst, no coverage, no float. ~0% institutional ownership, ~3.5M-share float, 12 nominal analysts and zero submitting estimates. Even if the gap should close, there is no marginal informed buyer positioned to close it.

Risks to thesis (tail, not bear case)

Discount never closes (value trap)High

The defining risk and the one the 0% prob-below-market does not price. Korean controlled holdcos trade 50–70% below NAV for decades. With no buyback, no NAV-closing event and an indifferent family, a minority holder can be right on value and collect nothing for years. This is a sell on realizable cash pending a named catalyst, not a clean buy.

Controlling-family related-party drainHigh

~53.5% family-held; the holdco is a succession vehicle run through private shell companies (S&S/SSY, BGH, CosM&M), with a history of dilutive discounted issuance to minorities. The 15% haircut prices the structural tax; a worse outcome justifies more.

Crown-jewel re-ratingMed

₩435B — the bulk of value — is one marked-to-market stake in Cosmax Inc. A China stumble (it is currently +20% Q1'26) or strong-KRW translation hit cuts the stake; the Monte Carlo varies it across the ₩155k–₩287k 52-week band.

Opco leverage / coverageMed

₩513.9B debt vs ₩107.2B cash, net interest cover ~1.0x, dividend cash-payout ~173% (funded by the associate pickup, not opco FCF). Synthetic credit ~Caa/CCC. Fragile if the equity-method income falters.

Suppressed-beta / illiquidityLow

~3.5M-share float and stale pricing suppress the regression β and distort the MC inputs. We address it by re-levering the sector β_u (1.78), but thin liquidity is a real friction for a minority position.

Soft domestic HFF demandLow

Cosmax NBT sits in a structurally soft Korean health-functional-food market; the export tailwind (US / SE-Asia) only partly offsets it. A drag on the opco — but the opco is a rounding error next to the stake.

10-year forecast

Opco revenue ₩664.6B → ₩867.1B over 10y (~3% fading to 2.5% terminal); operating margin lifts from 4.2% to 5.0%, staying below the 8.5% Food Processing global median. This chart is the ₩338B opco only — the ₩435B Cosmax Inc cross-holding and the 15% residual governance haircut, which drive the result, are outside it.

Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8 Y9 Y10 0 867.1B rev (KRW) 0% 5% op margin revenue FCFF op margin

Monte Carlo distribution

Across 1,000 correlated stress draws the p5 outcome is ₩27.4k/share, still +51% above the ₩18,120 price, and not one draw lands below the market — hence P(intrinsic < market) = 0.0%. But that 0% prices only the sampled input distributions (opco operating spread + the stake's 52-week range). It does not capture the real risk: that the structural NAV discount simply never closes. Read the band as a NAV ceiling, not a probability of profit.

p5 p25 p50 p75 p95 market 18120.00 14346.4 30941.9 47463.1 freq equity / share (KRW)

Mean ₩30910.33 ± ₩5615.63/sh, 1000 iterations (0 failed). P(intrinsic < market ₩18120.00) = 1.0%.

Cost of capital build
Risk-free rate 2.80%
Mature-market ERP 4.23%
Levered β 1.78
Weighted CRP 0.70%
Cost of equity 11.60%
Pre-tax cost of debt (synth Caa/CCC) 5.50%
D / V ~75%
WACC 5.95%
Full year-by-year DCF
Year Revenue Op mgn EBIT EBIT(1−t) Reinvest FCFF PV
1 ₩664.56B 4.27% ₩28.35B ₩21.26B ₩11.39B ₩9.88B ₩9.32B
2 ₩684.49B 4.35% ₩29.76B ₩22.32B ₩11.73B ₩10.59B ₩9.44B
3 ₩705.03B 4.43% ₩31.23B ₩23.42B ₩12.08B ₩11.34B ₩9.54B
4 ₩726.18B 4.51% ₩32.76B ₩24.57B ₩12.44B ₩12.13B ₩9.62B
5 ₩747.96B 4.59% ₩34.35B ₩25.76B ₩12.81B ₩12.95B ₩9.70B
6 ₩770.40B 4.67% ₩36.01B ₩26.91B ₩13.20B ₩13.71B ₩9.69B
7 ₩793.51B 4.76% ₩37.73B ₩28.09B ₩13.60B ₩14.49B ₩9.67B
8 ₩817.32B 4.84% ₩39.53B ₩29.32B ₩14.00B ₩15.31B ₩9.64B
9 ₩841.84B 4.92% ₩41.41B ₩30.59B ₩14.42B ₩16.17B ₩9.61B
10 ₩867.09B 5.00% ₩43.35B ₩31.91B ₩14.86B ₩17.05B ₩9.57B
Methodology & flags

Damodaran FCFF DCF, 10y explicit + perpetuity, in KRW. Risk-free 2.80% (local-currency government bond). Synthetic credit Caa/CCC. CRP 0.70% from revenue-weighted country mix × Damodaran 2026 CRPs. Terminal ROIC faded to 6.00%; 15% governance haircut applied post-DCF. Monte Carlo: 1000 iterations. Engine v1.0.0 · result: /Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/cosmaxbti-v2/output/2026-06-01-result.json

  • R&D cap: OFF · Lease cap: OFF · Failure: OFF · ESO: OFF
  • Governance haircut: 15% applied post-DCF (₩36195.31 > ₩30766.01)
  • Sensitivity tornado: not run