← back to picks

HQ Kaohsiung · Taiwan Founded 1971 · listed 1999 Reporting TWD Control Chen family ~31% Balance sheet net debt ~NT$3.75bn · D/E ~1.0 Listed TWSE · 0 analysts Valuation 2026-06-05 Damodaran FCFF · 1000-iter MC FCFF · Dark v3

Cheap on a peak PE, fair on normalized cash · +1.6% margin of safety at the base blend

WATCH — a real 10% yield wrapped around a fair-value DCF, with no floor

The screener flagged a +48% BUY; the normalized FCFF DCF says NT$32.81/share (pre-governance NT$36.45) against a NT$32.30 price — essentially fair value, and the Monte Carlo is a coin-flip at P(intrinsic < market) = 43.7% across 1,000 correlated draws. The +48% screen number was a FY2025 peak operating margin (24.6%), a floored suppressed β, and an un-modelled net-debt drag talking. Strip those out and the cheapness is gone: fair at the base blend (NT$32.81), −40% in the stress regime (NT$19.24), flattering only at the low-rate bookend (NT$43.47). Unlike net-cash Te Chang, LongDa is levered — ~NT$17/share of net debt and no balance-sheet floor to catch a margin miss. The one genuine attraction is a growing ~10% dividend: a yield play, not a value play. WATCH.

SectorHomebuilding (developer)Country mixTaiwan 100%β / MC σ1.46 levered · ±NT$5.65/sh (1000 runs)GovernanceChen-family control · 10% haircutQuality~11% ROIC ≈ WACC · net debt ~NT$3.75bn · no floorIncomeNT$3.3 cash + 0.5 stock · ~10% yield · 7y growth

Intrinsic / share

NT$32.81

post 10% gov · pre NT$36.45 · ≈ fair value

Market / share

NT$32.30

4 Jun 2026 close · TWSE

Margin of safety

+1.6%

vs intrinsic

Enterprise value

NT$11.74B

62.7% terminal

Cost of equity / debt

9.60% / 2.56%

β 1.46 · CRP 0.78%

Terminal ROIC / g

6.60% / 2.00%

spread ~45bp (ROIC 6.60% vs WACC 6.15%)

What it sells, where it sells

Operating segments

Property development & building salesResidential, Greater Kaohsiung — buy land, build, pre-sell, hand over. The lumpy, recognition-driven core~85%

Civil engineering & contractingTunnels, roads, bridges, public works — private + public-tender contracting~12%

Leasing + Japan hotel/restaurantRecurring-income tail (Motosu hotel near Mt Fuji, Kyoto restaurants) — immaterial to value~3%

LongDa reports largely consolidated, so the split is approximate, but FY2025 revenue is overwhelmingly the residential development core — the lumpy, completion-recognized line where the 24.6% operating-margin peak sits. Civil-engineering contracting and the small Japan hospitality tail add diversification but not value. The development margin is a handover peak on cheaply-acquired Kaohsiung land that anti-speculation policy and rising build costs will compress — explicitly not a run-rate to extrapolate.

Country mix (revenue-weighted CRP input)

🇹🇼Taiwan100%

Effectively a 100% Taiwan domestic property story, concentrated in Greater Kaohsiung. That pins the thesis to the Taiwan pre-sale cycle — now under the central bank's credit controls (luxury-loan caps, second-home tightening, hoarding tax 2.0) that the chairman has publicly blamed for unsold inventory — with no geographic diversification to cushion a slowdown.

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 +48% screener BUY is a triple artifact. The headline rested on (1) a FY2025 earnings peak (op margin 24.6%, gross 33%, net 20% — well above the 14–20% net the prior years ran, on cheaply-acquired land); (2) a suppressed/floored β (true 5Y regression β is −0.08, floored to 0.80 by the screen); and (3) no net-debt drag. Normalize the margin to a through-cycle 15%, re-lever β properly to 1.46, and subtract the real net debt, and intrinsic collapses to NT$32.81 — within ~2% of the NT$32.30 price.
Levered, asset-heavy, no floor — the key break from Te Chang. Total debt ~NT$6.8bn vs cash ~NT$3.1bn → net debt ~NT$3.75bn (~NT$17/share); D/E ~1.0–1.15; Altman-Z 2.28. Where Te Chang carried +NT$17/share of net cash as a hard floor under its WATCH, LongDa has the opposite — a net-debt drag and nothing to catch a margin miss. Inventory is NT$10.1bn (1.9× equity); asset turnover 0.37. A slow-turning, land-hungry balance sheet.
Fails the robustness gate. Cheap only at the low-rate bookend (NT$43.47, rf 1.13%); fair at the base blend (NT$32.81, rf 2.31%); −40% in the stress regime (NT$19.24, rf 4.30%). That is a pure rate-regime bet, not a robust buy — and the screen's own +6% stress MoS already warned the cushion was thin.
ROIC ≈ WACC → growth creates no value. Through-cycle ROC pins at ~6.6% ≈ the 6.15% WACC; reported ROE ~20% is flattered by leverage and the peak margin. A moat-less developer that earns its cost of capital is worth roughly invested capital, so the terminal perpetuity manufactures no excess returns — and 62.7% of EV already sits in that terminal block.
Controlled, but with a real and growing dividend. The Chen family controls ~31% via a web of investment vehicles (Da Jin 18.9% — ~12% of its stake pledged — Da Hong 9.4%, Feng Huang, Hong Ji…); 0 analysts; chronic stock-dividend dilution (FY25 0.5 stock div ≈ +5% shares) and a long convertible-bond history. Offsetting that: a genuine, 7-years-growing NT$3.3 cash + 0.5 stock dividend, ~10% yield, ~60% payout, that demonstrably reaches minorities — the one clean positive.

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 / peak reference

Value impact

01Revenue growth10-year path vs Homebuilding default

3% Y1-5Below the 5% screener default — peak/lumpy base into a cooling pre-sale market; the NT$26bn pipeline is visibility, not a higher run-rate; fades to 2.0% terminal

~5%Generic developer growth default

−NT$2

02Operating marginYear-10 target vs the FY25 peak

15.0%FY25 peak 24.6% (handover mix on cheap land) faded DOWN toward the through-cycle norm — above global Homebuilding 10.7% (Taiwan land economics), well below the peak

24.6% FY25The cyclical peak we refuse to extrapolate

−NT$8

03Sales-to-capitalReinvestment efficiency

0.55×Land/WIP-heavy developer (inventory 1.9× equity) — large capital per dollar of sales, so growth consumes land/WIP and creates little value (correctly)

1.27×Global Homebuilding median — LongDa is far more capital-intensive

−NT$3

04Terminal ROCYear 10+ return vs cost of capital

6.6%Engine-resolved ≈ terminal WACC — a moat-less developer's ROIC ≈ WACC, so terminal growth is value-NEUTRAL, not value-creative

≈ WACCNo durable excess returns

NT$0

05Cost of capital10y WACC vs the suppressed-β WACC

6.15%β 1.46 (global Homebuilding β_u 0.822 re-levered at D/E 0.96) · rf 2.31% blend · CRP 0.78%

~5%Suppressed 5Y regression β −0.08 floored to 0.80 → an artificially LOW WACC (the trap we rejected)

−NT$4

Why this ladder lands at fair value

Every honest override cuts the value the screener inflated. Fading the 24.6% peak margin to 15%, imposing the real capital intensity (0.55×), pinning terminal ROC at WACC, and rejecting the suppressed-β WACC together pull the FCFF to NT$32.81/share post-governance — within ~2% of the NT$32.30 price. There is no balance-sheet floor beneath it (net debt ~NT$3.75bn), so the only cushion is the ~10% dividend. WATCH, not BUY.

intrinsic NT$33

Our override Sector / peak reference Override adds value Override 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 engine inputs.

The 10-year story

LongDa is a 55-year-old Kaohsiung residential developer on a build-then-sell model: it buys land, builds, pre-sells, and recognizes the revenue in a lump when each project hands over. That makes earnings violently lumpy — FY2025 revenue fell 13% (because FY2024 carried a one-off land-sale gain) even as profit rose, and Q1 2026 revenue was down again. FY2025's 24.6% operating margin (33% gross, 20% net) is a peak: a high-margin handover mix on land bought cheaply years ago, into a market the chairman himself says is now choked by anti-speculation policy, credit controls, and rising build costs (the "soil-disposal" cost inflation he flagged twice on the earnings call). So we fade that margin to a through-cycle 15% — still above the global Homebuilding median because Taiwan land economics are favorable, but far below the peak — grow revenue a modest 3% off the base, and impose the heavy land-and-build capital intensity that is the real model. Pinning terminal returns at the cost of capital (growth that creates no value, because a moat-less developer earns roughly its WACC), the normalized cash is worth about NT$32.81 a share against a NT$32.30 price: fair value, not a bargain. The crucial difference from net-cash Te Chang — the other developer in this batch that screened cheap and converged to WATCH — is that LongDa is levered, ~NT$17 a share of net debt, with no balance-sheet floor if the margin reverts. What you are really buying is a genuine, seven-years-growing ~10% dividend; that is an income case, not a margin-of-safety case, and the stress regime (−40% at a 4.3% risk-free) shows how little protection the price leaves if rates normalize.

Two debates worth pressure-testing

A 6.5× PE with a 10% dividend yield — isn't that obviously cheap?

Our view: The 6.5× is on peak EPS. FY2025's 24.6% operating margin is a handover-mix peak on cheap land; normalize it to a through-cycle 15%, re-lever the suppressed β to 1.46, and subtract the ~NT$3.75bn net debt, and the intrinsic is NT$32.81 — basically the price. The 10% yield is real and the best reason to own it, but it is income, not a discount: there is no asset floor under a levered developer, and the dividend itself leans partly on the balance sheet in lean years.

The screener said +48%. Why is the answer WATCH?

Our view: The +48% was a triple artifact — peak margin, a floored suppressed β, and no net-debt drag. Fix all three and intrinsic falls to NT$32.81 (+1.6%), with the Monte Carlo a coin-flip at P(intrinsic < market) 43.7%. It also fails the robustness gate: −40% in the stress regime. Re-arm to BUY only if the through-cycle margin proves durably above 15%, or the property-policy regime clearly loosens.

CLAIM 01Revenue is lumpy and through-cycle, not a higher run-rate.rev growth 3% Y1-5 → 2.0% terminalBelow the 5% default: completion-recognized revenue swings hard (FY24 6.25bn on a land-sale gain, FY25 5.45bn, Q1-26 down again). The ~NT$26bn pipeline (NT$6bn inventory + NT$14.9bn 2027–30 + NT$5.57bn urban renewal) gives visibility, not a permanently higher base.

CLAIM 02Fade the 24.6% peak operating margin to ~15% through-cycle.target OPM 15% (from 24.6% FY25 peak)FY25's 24.6% OPM (33% gross) is a handover peak on cheaply-bought land; prior years ran 14–20% net. Anti-speculation policy, credit controls, and soil-disposal/aluminium-formwork cost inflation compress it. 15% sits above global Homebuilding 10.7% but well below the peak — normalization in the flows, not a caveat.

CLAIM 03Low sales-to-capital makes growth expensive — correctly.sales/capital 0.55× (Homebuilding 1.27×)A land-hungry developer ties up enormous capital per dollar of sales (inventory NT$10.1bn, 1.9× equity; asset turnover 0.37). Growth consumes land/WIP and creates little value — the reinvestment drag the screener omitted.

CLAIM 04Terminal ROC ≈ WACC — growth creates no value.stable ROC ≈ 6.6% ≈ terminal WACCA moat-less developer's ROIC tracks its cost of capital, so the terminal perpetuity manufactures no excess returns. With 62.7% of EV in that terminal block, pinning ROC at WACC is what keeps the low-rate-TWD perpetuity honest.

CLAIM 05Controlled and levered — but a real, growing ~10% dividend.gov haircut 10% · net debt ~NT$3.75bn · div NT$3.3+0.5Chen family ~31%, 0 analysts, chronic stock-dividend dilution, ~12% pledge on the largest vehicle — offset by a genuine 7-years-growing dividend (~10% yield, ~60% payout) that reaches minorities. Condition A (cash return) passes cleanly; the leverage means there is no NAV floor, so it's a yield case, not a value case.

Where we diverge from sell-side

We reject the screener's own +48% headline. It was a peak-margin × suppressed-β × no-net-debt artifact — the same shape as the Te Chang rebuild. Normalized intrinsic is NT$32.81 (+1.6%), not a +48% bargain.
No balance-sheet floor — the break from Te Chang. LongDa carries ~NT$3.75bn net debt (~NT$17/share), where Te Chang's WATCH at least rested on +NT$17/share of net cash. A margin miss here has nothing beneath it but the dividend.
Overvalued the moment rates normalize. Low-rate bookend NT$43.47 (+35%), base NT$32.81 (+2%), stress NT$19.24 (−40%). The price is justified only in the suppressed-rate regime — it fails the robustness gate.
The dividend is the thesis, not the DCF. A genuine, 7-years-growing ~10% yield (≈8% after Taiwan withholding) is the real reason to own it — an income play. But it partly leans on the balance sheet in lean years, so it is a yield to monitor, not a self-funding compounder.
Governance haircut 10%, matching Te Chang. Family control + chronic stock-dividend dilution + a ~12% pledge, balanced against a dividend that actually reaches minorities. Not higher (the cash is real); not lower (the dilution and leverage stack).

Two-sided case — bear anchors

Earnings revert from a peak. FY2025 op margin 24.6% / net 20% sits well above the 14–20% net the prior years ran. As the handover mix and cheap-land vintage roll off, reported EPS falls toward the through-cycle mid — the 6.5× PE does not discount that.
Policy is actively hostile. The chairman publicly blames anti-speculation policy, second-home credit controls and hoarding tax 2.0 for unsold inventory (房屋賣不出去); construction-cost inflation (soil disposal, aluminium formwork) squeezes margins. 100% domestic — no offset if the Taiwan cycle turns.
Levered into the downturn, with refinancing needs. Net debt ~NT$3.75bn, D/E ~1.0–1.15, Altman-Z 2.28; recurrent NT$2bn secured-bond programs roll bank debt. A property correction hits a levered developer's equity first, and there is no net-cash cushion.
Chronic dilution erodes per-share value. Annual stock dividends (~5% share growth) plus a long convertible-bond history quietly expand the count — the headline EPS and dividend per share are paid on a slowly growing base.
No catalyst to re-rate. No buyback-and-cancel, no asset-monetization plan, no analyst coverage to close the gap. Absent a margin surprise or a policy loosening, it stays a fairly-priced yield stock — dead money on price, paid to wait.

Risks to thesis (tail, not bear case)

Peak-margin reversionHigh

FY25's 24.6% OPM / 20% net is a handover peak on cheaply-acquired land; prior years ran 14–20% net. As the vintage and mix normalize, EPS falls toward the through-cycle mid — and the price has not discounted it. This is the dominant swing factor in the Monte Carlo.

Leverage / no floorHigh

Net debt ~NT$3.75bn (~NT$17/share), D/E ~1.0–1.15, Altman-Z 2.28. Unlike net-cash Te Chang there is no balance-sheet floor; a margin miss or property correction hits the equity directly, and recurrent bond programs must roll.

Taiwan property policyHigh

100% domestic. Central-bank credit controls (luxury-loan caps, second-home tightening, hoarding tax 2.0) have cooled the pre-sale market; the chairman blames them for unsold inventory. Construction-cost inflation pressures margins. No geographic offset.

Dividend sustainabilityMed

The ~10% yield is the core attraction, but in lean years it leans partly on reserves/debt rather than free cash, and ~21% non-resident withholding cuts the net. A cycle turn could force a cut — the income thesis is not bulletproof.

Rate-regime sensitivityMed

The 6.15% base WACC rests on a 2.31% TWD risk-free blend. Fair at base, the name is −40% in the stress regime (4.3% rf) — it fails the robustness gate, so a rate normalization removes what little margin exists.

Control & dilutionLow

Chen family ~31% via a web of vehicles (Da Jin ~12% pledged); chronic ~5%/yr stock-dividend dilution and a long convertible history. No disclosed abuse, but minorities have no lever and the per-share base quietly grows. Captured in the 10% governance haircut.

10-year forecast

Revenue NT$5.45B → NT$7.19B over 10y (3% Y1-5 fading to 2.0% terminal); operating margin faded from the FY25 24.6% peak to a through-cycle 15%. This is the normalized path behind the NT$32.81 intrinsic — the peak is a handover-mix high on cheap land, not a run-rate to extrapolate.

Monte Carlo distribution

The Monte Carlo spans NT$24.84 (p5) to NT$43.79 (p95) and lands a coin-flip across the NT$32.30 price — P(intrinsic < market) = 43.7%, median NT$33.16. The question is not whether it is cheap but whether the through-cycle margin holds: the upside draws are the ones where 15%+ margins persist, the downside draws are where policy and costs compress them. There is no asset value above the band to fall back on.

Mean NT$33.61 ± NT$5.65/sh, 1000 iterations (0 failed). P(intrinsic < market NT$32.30) = 43.7%.

Cost of capital build

Risk-free rate	2.31%
Mature-market ERP	4.47%
Levered β	1.46
Weighted CRP	0.78%
Cost of equity	9.60%
Pre-tax cost of debt (synth Aaa/AAA)	3.20%
D / V	~49%
WACC	6.15%

Full year-by-year DCF

Year	Revenue	Op mgn	EBIT	EBIT(1−t)	Reinvest	FCFF	PV
1	NT$5.62B	23.24%	NT$1.31B	NT$1.04B	NT$297M	NT$744M	NT$701M
2	NT$5.78B	21.87%	NT$1.26B	NT$1.01B	NT$306M	NT$703M	NT$624M
3	NT$5.96B	20.49%	NT$1.22B	NT$974M	NT$315M	NT$658M	NT$551M
4	NT$6.14B	19.12%	NT$1.17B	NT$936M	NT$325M	NT$611M	NT$481M
5	NT$6.32B	17.75%	NT$1.12B	NT$895M	NT$335M	NT$560M	NT$416M
6	NT$6.51B	16.37%	NT$1.07B	NT$851M	NT$345M	NT$506M	NT$354M
7	NT$6.71B	15.00%	NT$1.01B	NT$803M	NT$355M	NT$448M	NT$295M
8	NT$6.88B	15.00%	NT$1.03B	NT$825M	NT$325M	NT$500M	NT$310M
9	NT$7.04B	15.00%	NT$1.06B	NT$845M	NT$292M	NT$553M	NT$322M
10	NT$7.19B	15.00%	NT$1.08B	NT$862M	NT$256M	NT$606M	NT$331M

Methodology & flags

Damodaran FCFF DCF, 10y explicit + perpetuity, in TWD. Risk-free 2.31% (local-currency government bond). Synthetic credit Aaa/AAA. CRP 0.78% from revenue-weighted country mix × Damodaran 2026 CRPs. Terminal ROIC faded to 6.60%; 10% governance haircut applied post-DCF. Monte Carlo: 1000 iterations. Engine v1.0.0 · result: /Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/longda/output/2026-06-05-result.json

R&D cap: OFF · Lease cap: OFF · Failure: OFF · ESO: OFF
Governance haircut: 10% applied post-DCF (NT$36.45 > NT$32.81)
Sensitivity tornado: not run

FCFF DCF · dark theme v3. Generated 2026-06-05T01:51:42 from /Users/valentin/Documents/notes_jiliac_labs/Finance/Damodaran/valuations/longda/output/2026-06-05-result.json.