STL: small UPCoM real-estate name with deep historical losses and model-implied upside but very low confidence
Intrinsic value VND 1,348 vs market VND 1,000 => implied upside 34.8% (confidence: very_low).
Business Overview
Công ty Cổ phần Sông Đà - Thăng Long is a small-cap real estate company listed on UPCoM with 15,000,000 shares outstanding. The company sits in the Bất động sản sector and reports under VAS accounting; its asset base has declined from VND 5,170.3 bn in 2014 to VND 4,546.8 bn in 2016 according to the provided historical series. Institutional holders are significant: three funds/vehicles together own roughly 38.2% (ASEAN Deep Value Fund 14.3%, ASEAN Small Cap Fund 12.0%, Lucerne Enterprise Ltd 11.9%), leaving limited free float and some foreign room (foreign_room: 2,571,700.05 shares).
Investment Thesis
Valuation: our blended DCF/RNAV screening model produces an intrinsic per-share value of VND 1,348 (raw calibrated DCF VND 112,801.2 but isotonic calibration and blend weights yield the per-share figure), implying 34.8% upside to the current market price of VND 1,000. However model confidence is flagged as very_low and the model also raised sanity flags ("very_extreme_upside", "illiquid", "illiquid_upside_capped").
Operating track record: reported revenues and profits in the historical sample are weak and volatile — revenue was VND 293.8 bn in 2014 then fell to VND 49.0 bn in 2015 and recovered to VND 111.7 bn in 2016, while net profit remained deeply negative (e.g., VND -967.9 bn in 2014, VND -910.6 bn in 2015 and VND -146.7 bn in 2016). Latest reported EPS is VND -9,780 per share and reported book metrics (BVPS, ROE, ROA, margins) are all zero in the provided ratios table, highlighting very weak recent profitability and/or missing statutory equity recognition under the dataset.
Balance sheet and cash flow: model inputs show significant net debt (VND 2.6 trillion, per inputs), and operating cash-flow series are not provided. Earnings quality is middling at 55.0/100. Given the company’s small free float, UPCoM listing and illiquid trading (avg_volume_2w = 0), execution and marketability risk are material.
Investment case summary: the upside vs market is sizeable on the model outputs, but confidence is very low because of (1) weak historical performance and negative EPS, (2) sizeable net debt, (3) illiquidity and calibration/sanity flags in the valuation, and (4) concentrated institutional ownership which can limit trading and increase price volatility. The stock may warrant selective accumulation by patient, long-horizon investors willing to accept execution and liquidity risk, but the upside is not supported by consistent operating improvement in the reported historicals.
Valuation Commentary
Blend of a calibrated DCF (60%) and RNAV-style revaluation (40%) using an isotonic calibration; DCF uses base cash flow and a terminal growth of 3.5% with WACC 11.97%.
- Base cash flow (model input) VND 349,519,252,015 underpins forward free cash flows.
- WACC 11.97% and terminal growth 3.5% (terminal_g = 0.035) determine present value of the terminal value (TV accounts for ~67.48% of the DCF value).
- Net debt of approximately VND 2.6 trillion reduces equity value materially in the leverage-screened DCF.
- Blend weights: 60% DCF / 40% RNAV with RNAV revaluation factor 1.5; isotonic calibration was applied to map raw intrinsic to per-share value.
- Sanity flags: very_extreme_upside and illiquidity led to a calibrated intrinsic per-share of VND 1,348 and a model confidence downgrade to very_low.
The model-implied upside of 34.8% indicates material valuation dispersion versus the market price, but the very_low confidence and explicit sanity flags mean this upside should be treated cautiously. Key sensitivities are the base cash flow and the terminal assumptions; given illiquidity and weak historical earnings, confidence in the point estimate is low and outcomes may vary widely.
Bull vs Bear
- Model-implied intrinsic per share VND 1,348 implies 34.8% upside from current market VND 1,000.
- Terminal value is a large component of DCF (TV_pct = 0.6748), so a modest improvement in long-term cash generation could re-rate value materially.
- Institutional ownership is meaningful (top three holders ~38.2%), which could support stability if one or more turn constructive.
- Historical net losses: net profit was VND -967.9 bn in 2014 and VND -146.7 bn in 2016, and latest EPS is VND -9,780 per share, indicating continued operating struggles.
- Material net debt (approx. VND 2.6 trillion) increases refinancing and balance-sheet risk; interest coverage in model inputs is listed at 24.36 but debt weight is zero in WACC, suggesting data inconsistencies.
- Severe liquidity limitations (avg_volume_2w = 0; UPCoM listing) and model sanity flags ("illiquid", "very_extreme_upside") mean the market may not efficiently price any private improvement in fundamentals.
Sector Context
The company sits in the Vietnamese real-estate sector where VAS accounting, land-use-rights valuation and occasional state-driven revaluations can create reporting differences with listed peers. SBV credit growth quotas and affordability constraints continue to influence project funding and demand for residential/commercial developments. Within the peer set of 123 names, the sector median model-implied upside is 22.1%; STL's 34.8% is above that median but model confidence is lower than many peers. Top peer upside examples include NRC (55.5% upside, confidence low) and AGG (41.3% upside, confidence medium). Compared with those peers, STL's small equity base (15,000,000 shares) and UPCoM trading status increase liquidity and governance risk.
Risk Factors
- Profitability risk: trailing historical net profits are deeply negative (e.g., VND -967.9 bn in 2014; VND -146.7 bn in 2016) and latest EPS is VND -9,780, indicating limited evidence of a sustainable earnings turnaround.
- Balance-sheet/refinancing risk: net debt per model inputs is approximately VND 2.6 trillion, which is large relative to market cap implied by VND 1,000 price and 15,000,000 shares (market cap VND 15,000,000,000 = VND 15.0 bn), creating material leverage and solvency concerns.
- Liquidity and marketability: avg_volume_2w = 0 and trading limited to UPCoM; institutional concentration (~38.2% among top three holders) can exacerbate illiquidity and lead to abrupt price moves.
- Valuation-model risk: the DCF produced an extremely large raw intrinsic number that required calibration and flags ("very_extreme_upside"), indicating high model sensitivity to inputs and limited robustness.
- Data/coverage gaps: many standard ratios are zero or missing (ROE, ROA, margins, BVPS), complicating cross-checks and forensic analysis under VAS reporting.
- Regulatory and sector risks: real estate developers in Vietnam are exposed to changes in land-use-right valuation, local permitting, and macro measures such as SBV credit curbs that can quickly alter project financing availability.
Catalysts
- Receipt of audited financials or clarification of book equity that reduces reporting uncertainty and fills current data gaps.
- Asset monetization or land-use-right revaluation that materially reduces net debt or generates one-off cash inflows.
- Change in ownership or a strategic investor increasing free float / liquidity could narrow the market-to-model discount.
Forensic Assessment
No Beneish M-Score provided and the forensic summary contains no red flags in the input. That said, many key reported ratios are zero or missing and the model flagged sanity issues; combined with UPCoM listing and concentrated ownership, the practical consequence is uncertainty about earnings quality. Earnings_quality is middling at 55.0/100. In absence of explicit forensic red flags, primary concerns are data completeness and accounting visibility under VAS rather than documented manipulation signals.
Track Record
Model track record is short: 1 year of coverage (2026) with no reported hit_rate and an average model-implied upside of 34.84% across the observed period. With such limited historical performance the predictive reliability is unproven; readers should treat the model output as exploratory rather than a high-confidence call.
Written by a language model on 2026-08-11 from this page’s own model outputs and financial statements, and may quote figures from that date. Descriptive analysis, not investment advice — no buy, sell or hold recommendation is given or implied.