← back to picks

HQ Taipei · Taiwan Reporting TWD Listed TWSE · founded 1988 Control family ~37% · no pledging Valuation 2026-06-05 Damodaran FCFF · NAV cross-check FCFF · Dark v3

Cheap on peak EPS, fair on the assets · a levered, sub-WACC developer worth ~its risk-adjusted NAV, not 2.3× it

Trades at 0.53× book for a reason — roughly fair once you normalize the peak year and respect the leverage

The screen flagged +126% on FY2025 EPS of NT$3.19 — but that is a 9-year high from a one-off completion-handover year; through-cycle EPS is ~NT$1.5–1.75, putting the normalized P/E at ~11–13×, right at the Taiwan RE-sector average. A normalized FCFF DCF returns negative equity (−NT$28.7/sh) because ~NT$16bn net debt swamps an earnings-power enterprise value of only NT$6.6bn — which itself proves the point: Hong Pu earns ROIC ~2–4% versus a ~4.6% WACC even at peak, so it creates no going-concern value and is rationally worth less than book. The market prices it on NAV, not earnings. Triangulating normalized earnings power (~NT$15–19) against a downturn-adjusted NAV (~NT$21–27, book ceiling NT$38) gives fair value ~NT$18–25, mid ~NT$22 — at NT$20.10 the stock is roughly fair: a WATCH, not the 2.3× bargain the peak P/E advertised.

SectorReal Estate (Development)Country mixTaiwan 100%β / WACC1.45 levered (regression 0.02 rejected) · WACC 4.58%GovernanceFamily ~37% · no pledging/dilution · 10% haircutQualityROIC ~2–4% < WACC · net debt 2.4× mcapIncomeNT$2.0 cash div (~9.8%) · lumpy, not growing

Intrinsic / share

NT$22.00

NAV + earnings band NT$18–25 · post 10% gov

Market / share

NT$20.10

4 Jun 2026 · TWSE

Margin of safety

+9.5%

vs intrinsic

Enterprise value

NT$6.56B

66.6% terminal

Cost of equity / debt

9.59% / 2.80%

β 1.45 · CRP 0.78%

Terminal ROIC / g

4.00% / 2.00%

spread ~-58bp (ROIC 4.00% vs WACC 4.58%)

What it sells, where it sells

Operating segments

Residential / commercial developmentBuild-and-sell housing, offices and parking in Greater Taipei + Taoyuan; revenue booked at project completion/handover — hence the violent year-to-year lumpiness94%

Recurring rental (Mitsui leg)和苑三井 hotel built and leased whole to Mitsui Fudosan + 世貿 office — small but stable cash-yield assets6%

The donut is FY2025 whole-company revenue — NT$10.46B, a record completion year (+327% YoY), not a run-rate. A developer recognizes a whole project at handover, so a single year is meaningless: through-cycle recognition is ~NT$5–6B. The ~6% recurring-rental leg (the Mitsui hotel and offices) is the only smooth revenue; everything else is project-timing.

Country mix (revenue-weighted CRP input)

🇹🇼Taiwan100%

Entirely a Taiwan story — every project sits in Greater Taipei or Taoyuan. That concentrates the thesis on the domestic housing cycle, currently in its worst downturn in a decade under the central bank's seventh round of credit controls (record ~195k unsold new units nationwide), with no geographic diversification to cushion it.

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.

The +126% screen reading is a peak-EPS mirage. FY2025 EPS of NT$3.19 is a 9-year high from a one-off completion wave (宏普時代苑 + Garden Park 2 handovers; revenue NT$10.46B, +327% YoY). The 11-year EPS record is violently lumpy (4.41/3.21/2.35/1.34/1.83/1.84/−0.10/−0.16/0.69/0.61/3.19; mean ~1.75). Normalize and the P/E is ~11–13× = the Taiwan RE-sector average. 2026 is already a recognition trough (Q1 revenue ~NT$3.5B annualized).
The FCFF DCF prints negative equity — and that is the signal, not a glitch. Operating enterprise value is NT$6.56B, but ~NT$16.2bn net debt (debt 18.9B − cash 2.6B) leaves equity at −NT$10.6bn (−NT$28.7/sh). The market cap is +NT$6.69bn, which back-solves to a market-implied operating-asset value of ~NT$23.9bn — the ~NT$17bn gap is the NT$28.8bn land/inventory book (at cost) the market expects realized as asset sales, not capitalized as earnings.
Sub-WACC returns justify the 0.53× book discount. ROIC is ~2–4% against a ~4.6% WACC — even at the FY25 peak, ROIC only reaches ~4.4%. A firm that earns below its cost of capital is rationally worth less than book; the discount is correct pricing, not a mispricing to arbitrage.
Heavily levered into a severe, policy-driven downturn. Net debt is ~2.4× the market cap (D/V ~74%); inventory is NT$28.8B (76% of assets). The 7th credit-control wave (Sep-2024) pushed nationwide unsold new-home inventory to a record ~195k units; Q1-2026 new-case transactions −36% YoY, permits −30%, starts −37%. Moderate inventory write-downs would erase a large share of a thin equity sliver.
But not a value trap: real assets, presold pipeline, aligned owners. The NT$42.7B in-construction pipeline (recognized 2024–2031) is fully presold; a recurring Mitsui-rental leg and a deferred land bank back the book. Family owns ~37% via holding companies with zero share pledging and no dilution in 10+ years; the NT$2.0 FY25 dividend (~9.8% yield) pays you to wait, though it is lumpy and EPS-linked, not growing.

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

01Revenue growthNormalized run-rate vs RE-development median

2% (norm)Level normalized to ~NT$5.5B through-cycle; no real growth (land bank deferred)

~3–4%Global Real Estate (Development) cross-sector median

~NT$0

02Operating marginYear-5 target vs global RE-dev aggregate

12.0%Faded from FY25 ~15% peak; mgmt explicitly guides margins DOWN (fixed presale prices vs +25% build cost, carbon fee)

~3.5%Global RE-development EBIT-margin aggregate (incl. many low-margin developers)

+EV

03Sales-to-capitalReinvestment efficiency vs sector

0.35×Capital-intensive developer; huge land/WIP inventory per dollar of revenue (trough snapshot ~0.18×)

~0.41×Global Real Estate (Development) median

−EV

04Terminal growthYear 10+ steady state vs TWD risk-free ceiling

2.0%TWD long-run inflation; held ≤ the 2.31% risk-free (Damodaran ceiling)

2.31%risk-free ceiling (50/50 TWD blend); terminal g held below it

NT$0

05Cost of capital10y WACC; β rejects the suppressed regression

4.58%β 1.45 (Damodaran-global RE-dev βu 0.446 re-levered at D/E 2.8) · rf 2.31% · CRP 0.78% · D/V ~74%

~regression 0.025Y regression β — suppressed-trap garbage for an illiquid micro-cap; REJECTED

−EV

Why this ladder doesn’t set the price

Every reasonable override here still produces a negative equity value, because equity = enterprise value − ~NT$16bn net debt, and the earnings-power EV (NT$6.6bn) is below the net debt. The reason is structural: terminal ROC (~4.0%) sits below WACC (4.58%), so the business creates no value at the margin. The honest valuation is therefore not a DCF — it is NAV: book equity ~NT$38/sh, risk-adjusted for the downturn and the sub-WACC drag to a fair-value band of ~NT$18–25.

fair ~NT$22

Our override Sector median Lifts enterprise value Lowers enterprise 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

Hong Pu is a 1988-vintage Taiwan property developer, family-controlled, building and selling homes and offices in Greater Taipei and Taoyuan, with a small recurring-rental leg through a partnership with Japan's Mitsui Fudosan. Its 2025 earnings were a 9-year record because several large projects finished and handed over in the same year — but a developer books a whole project only at handover, so that NT$3.19 EPS is a spike, not a run-rate; normalized through-cycle earnings are roughly NT$1.5–1.75 a share, which makes the “cheap” 6× P/E an ordinary ~11–13× once you average the cycle. Run a full DCF and the equity value comes out negative, because the company carries about NT$16 billion of net debt against an earnings stream worth only ~NT$6.6 billion — which is the real tell: it earns a return on capital below its cost of capital even in a good year, so as a going concern it is worth less than its book value, and the market correctly prices it at about half of book. The value that does exist is in the assets — NT$28.8 billion of prime-Taipei land and buildings carried at cost — so this is a net-asset-value story, not a compounding one. Weighing normalized earnings power (~NT$15–19) against a downturn-adjusted asset value (~NT$21–27, with book at NT$38 the ceiling) lands fair value near NT$22, versus a NT$20 price. That is roughly fair, not a bargain — a WATCH, with the upside contingent on prime land holding its value through the worst Taiwan housing downturn in a decade.

Two debates worth pressure-testing

It trades at half of book with a fully-presold pipeline — isn’t that a screaming buy?

Our view: No. Half-of-book is the correct price for a developer that earns ~2–4% on capital against a ~4.6% cost of capital — sub-WACC firms belong below book (justified P/B ≈ ROIC/WACC). The presold pipeline removes demand risk on those units but not margin risk (fixed prices vs rising costs) or balance-sheet risk (net debt 2.4× mcap). You’re paying ~NT$20 for a risk-adjusted NAV of ~NT$22 — fair, with a cushion, not a steal.

Why publish a negative DCF number at all?

Our view: We don’t headline it — we use it as proof. The −NT$28.7 print shows there is no going-concern premium on top of the assets: the entire equity value is the realizable worth of the land book. That redirects the valuation to NAV, where the honest fair value (~NT$18–25) actually lives. A credit-control easing plus evidence the 2029–31 pipeline holds margin would flip this from WATCH to BUY.

CLAIM 01The +126% screen MoS is a peak-EPS artifact.FY25 EPS NT$3.19 (9-yr high) vs through-cycle ~NT$1.5–1.75A developer books whole projects at handover, so 2025’s completion wave is a spike. On normalized EPS the P/E is ~11–13× = the Taiwan RE-sector average — no real cheapness. Same windfall-EPS trap that turned Hua Eng into a SELL.

CLAIM 02Normalized FCFF equity is negative — a feature, not a bug.EV NT$6.56B − net debt ~NT$16.2B → equity −NT$10.6B (−NT$28.7/sh)Net debt swamps the earnings-power EV. It proves there is no going-concern premium above the assets — the equity value is purely the realizable worth of the land book.

CLAIM 03Sub-WACC returns justify the 0.53× book discount.ROIC ~2–4% < WACC 4.58% (peak ROIC only ~4.4%)A firm earning below its cost of capital is worth less than book; justified P/B ≈ ROIC/WACC ≈ 0.4–0.9. The market’s half-of-book price is rational, not a mispricing.

CLAIM 04It’s an NAV story: fair value ~NT$18–25, mid ~NT$22.earnings power NT$15–19 · risk-adj NAV NT$21–27 · book NT$38Triangulating normalized earnings power against a downturn-haircut NAV centers near NT$22 — ~+9% over the NT$20.10 price. A cushion, not a margin of safety worth a BUY.

CLAIM 05Levered into a record-inventory downturn — but aligned owners.net debt 2.4× mcap · family ~37%, no pledging/dilution · NT$2.0 divModerate inventory write-downs would erase much of a thin equity sliver — the key risk. Offsetting: presold pipeline, deferred land bank, a recurring Mitsui-rental leg, and a ~9.8% (lumpy) dividend that pays you to wait.

Where we diverge from sell-side

We reject the screen’s +126% headline. It anchors on a record completion-year EPS; normalized, Hong Pu is ~fairly valued, not a 2.3× bargain.
We value it on NAV, not the DCF. For a sub-WACC, heavily-levered asset-conversion developer the FCFF DCF prints negative equity and is the wrong primary lens — the assets, not the earnings stream, carry the value.
We reject the 0.02 regression beta. A suppressed-trap artifact for an illiquid micro-cap that fell 36% in a year; we use the Damodaran-global RE-development βu (0.446) re-levered at the real D/E → β 1.45, WACC 4.58%.
We normalize the base year, not annualize the peak. Through-cycle revenue ~NT$5.5B and margins faded toward the sector — against sell-side EPS estimates that lean on the 2025 completion wave.
We treat the 0.53× book discount as justified, not a free lunch. Sub-WACC returns + leverage + a record-inventory downturn explain the discount; the asset cushion is real but contingent.

Two-sided case — bear anchors

Leverage turns a moderate downturn into an equity wipe. Net debt is 2.4× the market cap and inventory is 76% of assets; even a ~10–15% write-down on a NT$28.8B land book at cost would erase a large share of the thin equity. Nationwide unsold inventory is at a record ~195k units and still climbing.
Margins are guided down, structurally. Presale prices are fixed but construction cost is +25% post-COVID, with a new carbon fee and soil-disposal spikes. Management itself says future gross margins should be viewed more conservatively — the 12% we model may prove generous.
Recognition is lumpy and bank-dependent. 2026 is a trough; the next big wave is 2029–31. Tightened mortgage lending already delayed handovers (Central Park 2 buyers), and any further deferral pushes earnings and cash further out.
The dividend won’t hold. The NT$2.0 FY25 payout (~9.8% yield) is EPS-linked off a peak year; it was NT$0.50 for FY23/24 and will shrink in the 2026–28 trough — the income leg is not a reliable carry.
100% domestic, no offset. The whole thesis rides the Taiwan housing cycle under the central bank’s seventh credit-control round; there is no geographic diversification if the domestic market deteriorates further.

Risks to thesis (tail, not bear case)

Inventory write-down on leverageHigh

Net debt 2.4× mcap; inventory NT$28.8B (76% of assets) carried at cost. A ~10–15% markdown in a record-inventory downturn (195k unsold units nationwide) would erase much of the thin equity sliver. This is the dominant risk.

Sub-WACC returns persistHigh

ROIC ~2–4% < WACC 4.6% even at peak. If margins fade as guided, the business keeps destroying value at the margin and the justified P/B stays near the low end (~0.4× book ≈ NT$15) — below today’s price.

Margin compressionMed

Fixed presale prices vs +25% construction cost, carbon fee, soil-disposal spikes. FY25 op margin 15% is a cyclical high; the 12% normalized assumption could prove optimistic, pulling NAV and earnings power lower.

Recognition timing / mortgage tighteningMed

Earnings cluster at handover (2026 trough, next wave 2029–31). Bank credit tightening already delayed buyer mortgages and handovers; further deferral pushes cash and earnings out and stresses refinancing of ~NT$15B of annually-rolling current debt.

Taiwan policy / housing cycleMed

100% domestic exposure to the central bank’s seventh credit-control round; transactions −36% YoY, permits −30%, starts −37%. No geographic offset if the downturn deepens.

Governance / family controlLow

Family ~37% via holding companies, 0 analysts. Mitigated by zero share pledging, no dilution in 10+ years, independent-majority board and a chairman/CEO split — better than the Taiwan family-holdco baseline. Priced via a 10% haircut.

10-year forecast

Normalized path: revenue ~NT$5.5B → NT$6.7B over 10y (~2% growth off a through-cycle base, NOT the FY25 NT$10.46B peak); op margin held at 12%, faded from the FY25 ~15% high. This stream is worth an EV of only NT$6.6B — below the ~NT$16bn net debt, which is why the FCFF equity is negative and the valuation defaults to NAV.

Monte Carlo distribution

Monte Carlo was not run: a probability band around a negative, tool-inappropriate point estimate would manufacture false precision. The thesis rests on NAV + earnings-power reconciliation (~NT$18–25, mid ~NT$22), not on the perpetual-FCFF spread.

⚠ Active diagnostic: stable_state.override_roc (0.0400) < WACC (0.0458); every dollar reinvested in stable state destroys value sales_to_capital_y1_5 (0.35) outside the typical [0.5, 5] band; verify the industry/business model sales_to_capital_y6_10 (0.40) outside the typical [0.5, 5] band; verify the industry/business model

Cost of capital build

Risk-free rate	2.31%
Mature-market ERP	4.47%
Levered β	1.45
Weighted CRP	0.78%
Cost of equity	9.59%
Pre-tax cost of debt (synth B1/B+)	3.50%
D / V	~74%
WACC	4.58%

Full year-by-year DCF

Year	Revenue	Op mgn	EBIT	EBIT(1−t)	Reinvest	FCFF	PV
1	NT$5.61B	12.00%	NT$673M	NT$561M	NT$314M	NT$247M	NT$236M
2	NT$5.72B	12.00%	NT$687M	NT$573M	NT$321M	NT$252M	NT$231M
3	NT$5.84B	12.00%	NT$700M	NT$584M	NT$327M	NT$257M	NT$225M
4	NT$5.95B	12.00%	NT$714M	NT$596M	NT$334M	NT$262M	NT$219M
5	NT$6.07B	12.00%	NT$729M	NT$608M	NT$340M	NT$268M	NT$214M
6	NT$6.19B	12.00%	NT$743M	NT$615M	NT$304M	NT$311M	NT$237M
7	NT$6.32B	12.00%	NT$758M	NT$622M	NT$310M	NT$312M	NT$226M
8	NT$6.44B	12.00%	NT$773M	NT$629M	NT$316M	NT$313M	NT$214M
9	NT$6.57B	12.00%	NT$789M	NT$636M	NT$322M	NT$314M	NT$202M
10	NT$6.70B	12.00%	NT$805M	NT$644M	NT$329M	NT$315M	NT$190M

Methodology & flags

Damodaran FCFF DCF, 10y explicit + perpetuity, in TWD. Risk-free 2.31% (local-currency government bond). Synthetic credit B1/B+. CRP 0.78% from revenue-weighted country mix × Damodaran 2026 CRPs. Terminal ROIC faded to 4.00%; 10% governance haircut applied post-DCF. Monte Carlo: — iterations. Engine v1.0.0 · result: /Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/hongpu/output/2026-06-05-result.json

R&D cap: OFF · Lease cap: OFF · Failure: OFF · ESO: OFF
Governance haircut: 10% applied post-DCF (NT$-31.85 > NT$-28.67)
Sensitivity tornado: not run

FCFF DCF · dark theme v3. Generated 2026-06-05T07:38:30 from /Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/hongpu/output/2026-06-05-result.json.