Fairly priced — the captive-supplier discount is already in the price · −8.2% margin of safety
Price sits at the top of the MC band — P(below) = 69.8%Intrinsic value ₩9,576/share (after a 10% governance haircut; pre-haircut ₩10,640) vs market ₩10,430 — a slim −8.2% at the regression β of 1.62. The Monte Carlo median is ₩9,296 and P(intrinsic<market)=69.8%. The upside (Metaplant operating leverage, battery-case re-rate, robot body-module option) is real but lives in the bull scenario at ₩23,124; the β-1.9 stress takes it to ₩8,968 (−14%). A hold, not a buy.
What it sells, where it sells
Operating segments
Hwashin is effectively one product line — chassis/body stampings for one customer group — so there is no segment diversification to fall back on. The only internal mix lever is the battery case (~8% of sales and growing), which is exactly the modest ASP uplift the base case credits inside the single segment; the robot body-module call option sits at sister-co Hwashin Precision and is explicitly excluded here.
Country mix (revenue-weighted CRP input)
The mix follows Hyundai Motor Group's global build footprint, not Hwashin's own choices: US revenue (₩704bn) overtook Korea (₩698bn) in FY2025 as the Georgia/Metaplant ramp landed. It is a production-geography map of HMG, so the blended country-risk premium (~0.95%) is a derivative of where HMG chooses to assemble — the India/Brazil tail carries most of the EM CRP.
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.
- A single-customer, single-product Tier-1 supplier. ~95-96% of revenue is Hyundai Motor Group (FY2025 audited Note 4: the top two HMG customers alone are 76.1% of consolidated sales). Hwashin builds chassis, sub-frames and body stampings at plants physically co-located with HMG assembly, holding the #1 internal chassis share (~57%) and 100% chassis share at HMG's newest US/India/Brazil plants.
- The market prices it as a price-taker, at ~4-5x forward earnings. No pricing power against a single OEM that extracts ASP price-downs as models age; operating margin has lived in a thin 4-6% top-of-band / 0-2% trough band with outright net-loss years in 2015, 2017, 2018 and 2020. FY2025 OPM 5.22% is already near the historical ceiling. P/B ~0.63x.
- FCF-negative through the overseas build-out. PP&E is ~45% of total assets; the Georgia/Metaplant (~₩210-240bn) and Yeongcheon battery-case plants drove free cash flow to roughly −₩96bn (FY24) and −₩146bn (FY25). Net debt ≈ ₩478bn; the cheapness reflects leverage and cash burn during the build.
- A real but thin and growing dividend; no buyback-and-cancellation. DPS ₩50 (2020) → ₩170 (2025), ~3.4x in five years — the direction test passes — but the policy is only ~7-9% of consolidated net income, so the magnitude test fails. Share count is stable at 34.92M (treasury 0.17M); no dilution, no cancellation.
- A chaebol-style unlisted-affiliate layer is the governance concern. Control runs Jeong Seo-jin → Global Auto Trading (unlisted, 16.43%) → Hwashin, alongside Hwashin Precision (listed KOSDAQ) and unlisted World / International Auto Trading distribution vehicles. Mitigants: no KFTC tunnelling fine, succession complete, 60% independent-director board with an all-outside audit committee.
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 engine inputs.
The 10-year story
Hwashin is a chassis-and-body stamping supplier whose value is a derivative of one customer's build plan: about 95-96% of sales go to Hyundai Motor Group, and the company physically co-locates its plants with HMG's. That makes it an incumbent with monopoly chassis positions at HMG's newest US, India and Brazil plants, but also a price-taker with no leverage over a single OEM that extracts price-downs as models age. The base case treats every input at or just above Hwashin's own through-cycle record: revenue grows about 3% a year as HMG volume compounds, from ₩1.96tn to roughly ₩2.64tn over ten years; the operating margin recovers from the current Georgia/tariff trough to about 5.0% — below the FY2025 record, because there is no pricing power to support more; and reinvestment stays capital-intensive at a 1.25 sales-to-capital, reflecting a balance sheet that is 45% PP&E and free-cash-flow negative through the overseas build-out. Crucially, the terminal return on capital is pinned just above the cost of capital, because Hwashin's ROIC has cleared 11% only twice in its entire history and otherwise tracks WACC. Putting the customer-concentration and low-return economics into the flows like this — rather than bolting on a flat discount — the intrinsic value lands at about ₩9,576 after a 10% governance haircut for the family-affiliate leakage, within roughly 8% of the ₩10,430 market price. This is a fairly-priced operating business, not a value trap and not a bargain.
Two debates worth pressure-testing
- We cap the margin at 5.0%, below the FY25 record — the sell-side extrapolates the recovery. Samsung Securities' BUY ₩16,000 models >6% margin on Metaplant operating leverage. We refuse to credit margin expansion beyond a historical ceiling that has never held, and put the captive-OEM extraction in the flows.
- Terminal ROC pinned at WACC, not above it. Where a bull would compound an excess-return franchise, Hwashin's ROIC cleared 11% only twice; we model a value-creation-neutral steady state, the single most important grounding fact in the file.
- The Metaplant/battery/robot upside is a scenario, not the base. It is real option value (₩23,124 bull), but it has not shown up in the numbers — Georgia was loss-making in FY25 and 1Q26 OP fell ~44%. We will not pull it forward.
- Governance modeled in the flows, with a thin 10% residual. Not a bolted-on flat “Korea discount” (the Cosmax-BTI-v1 anti-pattern); the residual prices only the unlisted-affiliate leakage and thin payout the flows cannot capture.
- The verdict is β-conditional. At the regression-confirmed β 1.62 the gap is −8%; at a β-1.9 stress it widens to −14%. Unlike a clean buy, Hwashin does not open a margin of safety even at its central case — so we say hold, not buy.
- HMG volume softens and tariffs persist. One customer, one geography-mix bet. A US tariff drag plus an HMG build-rate cut would hit revenue and the thin margin together, with no second segment to cushion it.
- Georgia stays loss-making. The Metaplant plant lost ₩11.5bn in FY25 and the loss widened to −₩7.2bn in 1Q26. If utilization recovery slips past 2027, the depreciation drag keeps consolidated margin below the 5% base assumption.
- β / rate sensitivity. ~137% of enterprise value sits in the terminal block and explicit-period FCFF is thin against heavy capex, so the discount rate is load-bearing. A β-1.9 stress alone takes intrinsic to ₩8,968 (MoS −14%); a higher KRW risk-free compounds it.
- Re-lever into the next plant. Net debt is already ~₩478bn after Georgia and Yeongcheon. A further debt-funded India/Brazil expansion before the current plants reach utilization would push leverage and cash burn higher through another build cycle.
- Affiliate leakage proves larger than priced. The unlisted Global / World / International Auto Trading distribution vehicles are a reported-small but opaque channel; if intra-group distribution margins are larger than disclosed, the 10% residual understates the leakage to minorities.
Risks to thesis (tail, not bear case)
~95-96% of revenue is Hyundai/Kia. A volume cut, a price-down cycle, or an HMG sourcing shift hits the whole business at once, with no second segment or customer to absorb it. This is the structural reason the market caps the multiple at ~4-5x.
~137% of EV is in the terminal block and explicit FCFF is thin against capex. At the regression β 1.62 the gap is −8%; a β-1.9 stress takes intrinsic to ₩8,968 (−14%). The verdict is conditional on the cost of capital — the defining reason for the hold.
The US plant was loss-making in FY25 (−₩11.5bn) and 1Q26 (−₩7.2bn, widening). Recovery depends on Kia Sportage HEV (May 2026) and Hyundai HEV (2027) utilization; a slip keeps the depreciation drag on consolidated margin past the modeled trough.
Free cash flow was roughly −₩96bn (FY24) and −₩146bn (FY25) on overseas plant build-out; net debt ~₩478bn. Another debt-funded plant before current ones reach utilization extends the burn and the leverage.
Operating margin has fallen to 0-2% with outright net losses (2015/17/18/20). If HMG accelerates price-downs or US tariffs stick, the through-cycle 5% base assumption is optimistic, not conservative.
Control runs through unlisted Global / World / International Auto Trading distribution vehicles; payout is only ~7-9% of NI with no buyback-and-cancellation. Mitigated by no KFTC fine, completed succession and a 60% independent board — priced via the 10% haircut.
10-year forecast
Revenue ₩1.96tn → ₩2.64tn over 10y (~3% HMG-volume-driven CAGR, fading to a 2.5% terminal). Operating margin recovers from the Georgia/tariff trough to ~5.0% by year 5 — below the FY25 record 5.22%, because there is no pricing power to support more. FCFF stays thin against the capital-intensive 1.25 sales-to-capital.
Monte Carlo distribution
The Monte Carlo median (₩9,296) sits ~11% below today's ₩10,430, and P(intrinsic < market) = 69.8% — roughly seven correlated stress draws in ten land below the price. The distribution says fairly-to-slightly-richly valued, not cheap.
Mean ₩9346.56 ± ₩1990.78/sh, 1000 iterations (0 failed). P(intrinsic < market ₩10430.00) = 69.8%.
Two β regimes
For a name with ~137% of enterprise value in the terminal block, the cost-of-capital input is load-bearing. The 5Y regression β of 1.62 is confirmed by three independent sources and is the decision number; we also report a β-1.9 stress sibling to show the verdict does not flip to a buy even at the central case — which is why this is a hold.
| Regime | Levered β | Ke | WACC | Intrinsic (post-gov) | vs ₩10,430 |
|---|---|---|---|---|---|
| Base (regression β) | 1.62 | 12.22% | 6.87% | ₩9,576 | −8.2% |
| Stress (β 1.90) | 1.90 | ~13.5% | ~7.4% | ₩8,968 | −14.0% |
Cost of capital build
| Risk-free rate | 3.82% |
| Mature-market ERP | 4.59% |
| Levered β | 1.62 |
| Weighted CRP | 0.95% |
| Cost of equity | 12.22% |
| Pre-tax cost of debt (synth A3/A-) | 5.50% |
| D / V | ~66% |
| WACC | 6.87% |
Full year-by-year DCF
| Year | Revenue | Op mgn | EBIT | EBIT(1−t) | Reinvest | FCFF | PV |
|---|---|---|---|---|---|---|---|
| 1 | ₩2.02T | 5.18% | ₩104.64B | ₩84.96B | ₩47.26B | ₩37.71B | ₩35.28B |
| 2 | ₩2.08T | 5.13% | ₩106.87B | ₩86.78B | ₩48.68B | ₩38.10B | ₩33.35B |
| 3 | ₩2.15T | 5.09% | ₩109.14B | ₩88.62B | ₩50.15B | ₩38.48B | ₩31.52B |
| 4 | ₩2.21T | 5.04% | ₩111.46B | ₩90.50B | ₩51.66B | ₩38.85B | ₩29.78B |
| 5 | ₩2.28T | 5.00% | ₩113.81B | ₩92.41B | ₩53.21B | ₩39.20B | ₩28.12B |
| 6 | ₩2.34T | 5.00% | ₩117.23B | ₩93.93B | ₩54.81B | ₩39.12B | ₩26.19B |
| 7 | ₩2.42T | 5.00% | ₩120.76B | ₩95.45B | ₩56.46B | ₩38.99B | ₩24.31B |
| 8 | ₩2.49T | 5.00% | ₩124.40B | ₩96.98B | ₩58.16B | ₩38.82B | ₩22.49B |
| 9 | ₩2.56T | 5.00% | ₩128.14B | ₩98.52B | ₩59.91B | ₩38.60B | ₩20.73B |
| 10 | ₩2.64T | 5.00% | ₩132.00B | ₩100.06B | ₩61.72B | ₩38.34B | ₩19.04B |
Methodology & flags
Damodaran FCFF DCF, 10y explicit + perpetuity, in KRW. Risk-free
3.82% (local-currency government bond). Synthetic credit A3/A-. CRP
0.95% from revenue-weighted country mix × Damodaran 2026 CRPs.
Terminal ROIC faded to 6.90%; 10% governance haircut applied
post-DCF. Monte Carlo: 1000 iterations. Engine v1.0.0 · result:
/Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/hwashin/output/2026-06-01-result.json
- R&D cap: OFF · Lease cap: OFF · Failure: OFF · ESO: OFF
- Governance haircut: 10% applied post-DCF (₩10639.70 > ₩9575.73)
- Sensitivity tornado: not run