Cơ khí An Giang (CKA): micro-cap cyclical with net cash but weak earnings quality and limited liquidity
Intrinsic value VND 64,956 vs market VND 67,800 — implied downside of 4.2% (model confidence: very_low).
Business Overview
Công ty Cổ phần Cơ khí An Giang (CKA) is a UPCOM-listed heavy-industry / mechanical engineering company serving agricultural and industrial equipment segments in Vietnam. It is classified in 'Công nghiệp nặng' and exhibits a small free float and low trading liquidity (avg volume 3,656 shares over 2 weeks; 1y high equals current price). The largest shareholder is the state-linked Tổng Công ty Máy động lực và Máy nông nghiệp Việt Nam with 47.41% ownership, leaving no foreign room (foreign_room 0.0%).
Investment Thesis
CKA combines some balance-sheet positives with pronounced earnings-quality and liquidity concerns. On the positive side, the model reports net_debt = -16047059056 (net cash), and the company generated VND 65.2 bn net profit in 2025 after a volatile prior two years (VND 41.2 bn in 2023; VND 23.8 bn in 2024). Profitability metrics look superficially strong: ROE 49.8% and ROA 33.8% are high and EPS is VND 19,840.6, while P/E is low at 3.4x and P/B 1.6x, implying valuation tailwinds if earnings are sustainable.
The countervailing factors are material. Earnings quality is weak (score 15.6/100) with a Beneish M-Score of -1.1655 (forensic risk_level: moderate) and explicit red flags around revenue recognition and possible aggressive accounting. Operating margin dynamics are unstable (EBIT margin -0.2% latest) and revenue has fallen over three years from VND 228.5 bn in 2023 to VND 162.4 bn in 2025. Trading is thin (UPCOM listing, average 2-week volume 3,656) and foreign ownership is closed, limiting liquidity and institutional demand. Given the model's intrinsic value of VND 64,956 per share versus a market price of VND 67,800 (implied -4.2%), and the model confidence flagged as very_low with sanity_flags including low_liquidity, low_earnings_quality, manipulation_risk, the implied margin for error is small versus execution and forensic risks.
Valuation upside is small and model confidence is very low; combined with concentrated SOE ownership (47.41%) and the forensic concerns, the risk/reward is unattractive for new capital at current prices.
Valuation Commentary
EV/EBITDA mid-cycle approach: we apply a fair EV/EBITDA multiple to a mid-cycle EBITDA to derive enterprise value and back out a per-share intrinsic value.
- mid_cycle_ebitda = 3,529,301,187 (source: own_median, years_of_data = 7)
- fair_ev_ebitda = 34.75 (source: own_history) vs sector_ev_ebitda = 9.14
- net_debt = -16047059056 (reported in model inputs) which increases equity value
- calibration: raw_intrinsic_value = 42,203.2 then isotonic calibration produced intrinsic_value = 64,956
- confidence downgraded to very_low due to low liquidity and earnings-quality/manipulation flags
The model implies a minor downside of 4.2% from current price but with very_low confidence. Given the large gap between the applied fair EV/EBITDA (34.75) and the sector median (9.14), plus calibration adjustments and forensic red flags, the intrinsic estimate is highly sensitive to assumptions. We place limited confidence in the precision of the intrinsic value and emphasize the forensic and liquidity caveats.
Bull vs Bear
- Net cash reported in model inputs (net_debt = -16047059056) supports higher equity valuation and reduces solvency risk.
- Strong trailing profitability metrics: ROE 49.8% and ROA 33.8% with EPS of VND 19,840.6 could justify low P/E if sustainable (P/E 3.4x).
- 2025 recovery in net profit to VND 65.2 bn after prior-year weakness shows potential for earnings cyclicality to reverse positively.
- Forensic flags: Beneish M-Score -1.1655 (risk_level: moderate) and earnings_quality 15.6/100 with revenue recognition score 0/100 indicate aggressive accounting and low earnings reliability.
- Revenue decline three-year trend from VND 228.5 bn (2023) to VND 162.4 bn (2025) and near-zero operating margin (EBIT margin -0.2%) imply fragile operational performance.
- Low liquidity (UPCOM listing, avg_volume_2w 3,656) and zero foreign room (foreign_room 0.0%) constrain re-rating potential and secondary-market exits for investors.
- High shareholder concentration: state-linked holder at 47.41% may limit free-float catalysts and leads to strategic, not market-driven, corporate actions.
Sector Context
CKA sits in the cyclical heavy industry segment where domestic demand for machinery and agricultural engines is tied to capex cycles and commodity prices. Sector EV/EBITDA median in the dataset is 9.14x, but CKA's calibrated fair EV/EBITDA is much higher at 34.75x according to the model’s history-derived inputs — a divergence that may reflect model idiosyncrasies or firm-specific asset valuation differences. UPCOM-listed industrials often suffer low liquidity and wider bid-ask spreads; VAS accounting and state ownership patterns can also obscure earnings comparability versus listed peers. SBV macro credit quotas are less directly relevant here, but broader capex/credit cycles influence end-market demand for heavy equipment.
Risk Factors
- Forensic/accounting risk: Beneish M-Score -1.1655 (> -1.78 threshold) and earnings_quality 15.6/100 highlight elevated risk of aggressive accounting, especially in revenue recognition (score 0/100).
- Liquidity and marketability: UPCOM listing, avg_volume_2w 3,656, and 1y high equal to current price point to extremely low trading liquidity; large positions may be difficult to exit.
- Concentrated ownership: state-linked holder owns 47.41%, which can limit corporate governance transparency, distributions, or strategic decisions aligned with minority shareholders.
- Volatile topline: revenue declined from VND 228.5 bn (2023) to VND 162.4 bn (2025), increasing execution risk for any valuation premia.
- Model sensitivity: valuation depends on a high fair_ev_ebitda (34.75) vs sector median (9.14) and isotonic calibration — small changes produce large moves in intrinsic value given thin free float.
- Regulatory/reporting differences: VAS accounting treatments and disclosure on UPCOM can lag HOSE/ HNX peers, reducing comparability and forensic clarity.
- Foreign ownership closed (foreign_room 0.0%): eliminates a potential investor base that could support re-rating or liquidity.
Catalysts
- Publication of audited annual report or auditor commentary that addresses revenue recognition and restores earnings-quality confidence.
- Material contract wins or order-book disclosures that reverse the three-year revenue decline and provide visible revenue visibility.
- Corporate actions increasing free float or improving governance (partial divestment by the state shareholder).
- A re-listing or transfer to HNX/HOSE that materially improves liquidity and investor coverage.
Forensic Assessment
Forensic signals are the primary concern. The Beneish M-Score of -1.1655 sits above the -1.78 manipulation threshold and the report flags an 85th percentile relative position, indicating a higher likelihood of aggressive accounting practices versus peers. The earnings_quality score of 15.6/100 — and a revenue recognition sub-score of 0/100 — specifically call into question the sustainability and transparency of reported revenue and profit. Offsetting this, the Altman Z-Score (reported in the forensic summary as 4.41) suggests low near-term bankruptcy risk, and a Piotroski F-Score of 4/9 is neutral. Overall, forensic assessment weighs heavily negative and is the dominant margin-of-safety concern for valuation.
Track Record
Model track record spans 9 years (first_year 2018, last_year 2026) with a hit_rate of 0.5, implying the model has been directionally correct roughly 50% of the time historically. Historical average upside when correct has been large (avg_upside_pct 99.8478), but the 50% hit rate suggests limited predictive precision. Given the very_low confidence on the current valuation and the company’s forensic flags, historical performance should be treated as a weak guide rather than reassurance.
Written by a language model on 2026-08-10 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.