Skip to content

CIO VERDICT — THG (Thonburi Healthcare Group PCL)

1. The call

  • Rating: HOLD (lean avoid)
  • 2-year target price: ฿7.80 (vs current ฿7.30, implies +6.8% price return; ~+8% total)
  • Suggested position size: 0% of book now; 0.4% maximum if entry conditions trigger
  • Time horizon: 18–24 months

Why: Six of seven specialists triangulate to a fair value of ฿7–9 — i.e., the stock is roughly where it should be. There is no edge worth paying for at ฿7.30. The bull case (RAM-driven margin convergence to peers) is real but unproven and capped by a 49.99% whitewash structure; the bear case (Boon-case contagion, big-bath FY24 unwind, further dilution) is live and quantifiable. With Auditor at C, Forensic at D, and Prosecutor at D, our risk gates explicitly forbid a conviction BUY regardless of narrative appeal. We wait for a setup — either ฿6.50 with a base, or a ฿8.65 reclaim with operational evidence.

2. How the specialists voted

Persona Verdict 2y target Key insight
Auditor Grade C n/a Prices clean; SET shareholders + filings 404'd; Stocktwits is wrong company (Hanover) — discard. Treat all line-items as unverified.
Trader No trade / Pass base ฿7–9; bull ฿14–16 Knife on expanding down-volume; no buyer defending; ฿8.65 placement = overhead supply. Wait for ฿6.50 base or ฿8.65 reclaim.
Analyst HOLD midpoint ฿7.50 (range ฿6–9) Recap was a balance-sheet fix, not an earnings turnaround. Revenue still declining (−4% FY25). ฿30 is fantasy on 2y view.
PI Caution qualitative CEO invisible 15 months post-takeover; 12 co-defendants unnamed = unexploded ordnance; 49.99% structure caps re-rating.
Prosecutor Grade D qualitative; −25 to −40% legal discount Three live files: issuer disclosure failures, AMLO tracing, whitewash challenge. Market is pricing legal risk approximately right.
Forensic Grade D (provisional) qualitative; 20–30% disclosure discount FY24 −฿1.76bn → FY25 +฿96m on declining revenue = big-bath risk. Cannot compute Beneish/Altman — refuses to fabricate.
Quant Wide CI 24m median ฿7–9; 90% CI ฿3–฿16 GBM drift=0 → ฿7.30; mean-revert to PP → ฿8.40; naive EPS×P/E → ฿0.86–฿5.40. ฿30 is >2σ outside any band.
Geopolitical Headwind moderate −5% overlay No tailwinds; ESG blacklist until 2027; whitewash re-examination tail risk; THB second-order.

3. Where they converge (the well-supported view)

  1. Fair value is ~฿7–9. Analyst midpoint ฿7.50, Quant 24m median ฿7–9, Trader base case ฿7–9, mean-reversion model ฿8.40. Spot ฿7.30 sits at the low end. The market is not grossly mispricing this.
  2. The ฿30 owner narrative is not a 2-year case. Every specialist independently dismisses it. It requires 30–35× FY25 earnings and a multiple re-rating that the legal/governance overhang explicitly prevents.
  3. The FY25 swing to profit was a balance-sheet event, not an earnings recovery. Revenue is shrinking; >฿300m interest savings did the work. Operating leverage not yet visible.
  4. Governance gates are tripped. Auditor C + Forensic D + Prosecutor D — the legal/disclosure complex is unresolved and unverifiable from the package.
  5. 49.99% whitewash structure caps upside. No tender premium, ever, on current terms. RAM's own cost basis (฿8.65) is a soft ceiling for 12+ months.

4. Where they diverge (the real questions)

Disagreement #1: Is FY25 EPS extrapolable? - Quant's naive P/E model says fair value is ฿0.86–฿5.40 if you capitalise FY25 EPS at peer multiples. - Analyst extrapolates to FY27E EPS ฿0.30 and gets ฿6–9. - My call: side with Analyst on directional EPS growth (full-year interest savings + modest synergy), but apply Forensic's 10% quality-of-earnings discount. FY25 is a poor base; FY27E EPS ฿0.22–0.30 is the right anchor, not ฿0.054.

Disagreement #2: How much does legal tail risk discount fair value? - Prosecutor: 25–40% off clean-peer multiple. - Forensic: 20–30% off until founder-related exposures are written down. - Geopolitical: −5% probability-weighted. - My call: side with Prosecutor/Forensic at the issuer level (~25% discount embedded in any peer-multiple derivation). Geopolitical's −5% is too generous because it doesn't include the AMLO tracing tail.

Disagreement #3: Is the chart entry-able? - Trader: no trade at ฿7.30; wait for ฿6.50 or ฿8.65 reclaim. - Analyst: fairly priced, HOLD. - My call: Trader wins on tactical entry. Closing on the low, expanding down-volume, no support visible. Even if Analyst's fair value is right, paying full price for a fairly-priced asset with negative momentum is poor risk management.

Disagreement #4: Will RAM defend ฿8.65? - PI: RAM is "hostage" at ฿8.65, not buyer. - Trader: same read — ฿8.65 is overhead supply, not support. - My call: ฿8.65 is a ceiling, not a floor, for 12+ months. RAM has no incentive to buy more above 49.99% (consolidation risk) and no incentive to sell below cost. They sit.

5. The 2-year forecast — my synthesized number

  • Bull case (20%): ฿11.50 — requires (a) FY26 net margin to 3%+ (~฿300m NI), (b) Boon case formally resolved or fully ring-fenced from THG, (c) named CEO with credible operational plan, (d) one clean audit cycle. Re-rate to ~12× FY27E EPS ฿0.30 = ฿11.
  • Base case (55%): ฿7.80 — RAM digestion continues, FY26 NI ฿250–400m, no Boon escalation but no clean resolution either, ESG overhang persists. Stock chops ฿6.50–฿9.00. Mean reverts toward PP anchor over time.
  • Bear case (25%): ฿4.50 — Either (a) FY26 swings back to loss, (b) Boon-case asset tracing reaches THG operating entities, (c) further dilution, or (d) FY24 big-bath unwinds (impairments not yet taken). Magnet at ฿5.00 rights price, then ฿4-handle.

Expected value: 0.20×฿11.50 + 0.55×฿7.80 + 0.25×฿4.50 = ฿2.30 + ฿4.29 + ฿1.13 = ฿7.72

vs current ฿7.30 → expected return +5.8% over 24 months, or ~+2.9%/yr. That's below the Thai risk-free rate. This is not an investment; it's a coin flip with a small positive drift.

6. Position-sizing math

  • Edge: EV ฿7.72 vs spot ฿7.30 = +5.8% over 24 months ≈ +2.9%/yr
  • Vol-adjusted edge: edge/σ = 2.9% / 40% = 0.07 (Sharpe-equivalent — terrible)
  • Kelly fraction (full): edge/variance ≈ 5.8% / (40%)² = 5.8% / 16% = ~36% (raw Kelly)
  • But raw Kelly is wrong here because it assumes normal returns. The Quant's 90% CI of ฿3–฿16 and three D-grade fat-tail risks mean true variance is materially higher. Adjusting σ to 55% to reflect fat tails: Kelly ≈ 5.8% / 30% = ~19%
  • Practical Kelly (×0.25 for parameter uncertainty): ~5%
  • Caps applied:
  • Data-integrity cap (Auditor C): halve → 2.5%
  • Tail-risk cap (Prosecutor D + Forensic D): halve again → 1.25%
  • Liquidity cap (sub-฿30M ADV, can't exit in size): cap at 0.5%
  • Conviction cap (low — specialists triangulate to "no edge"): cap at 0.5%
  • Final practical position size: 0.5% maximum, only on a tactical entry trigger. At ฿7.30 today: 0%.

7. Risk gates

Tail-risk gate: TRIPPED. Prosecutor D + Forensic D. Per the framework, this forces HOLD/SELL regardless of bottom-up enthusiasm. The Prosecutor explicitly identified three open files (issuer disclosure, AMLO tracing, whitewash challenge) that could each independently produce a 10–20% drawdown. No conviction BUY possible.

Data-integrity gate: TRIPPED. Auditor grade C with Low confidence on social/SET shareholder/filings data. SET shareholders and filings pages 404'd. I cannot verify the RAM 49.99% stake, the Vanasin 4.2% residual, post-recap net debt, related-party notes, share-pledge status, auditor identity, or the FY25 line items. This alone caps position size and forbids high-conviction targets. I have applied the DO-NOT-FABRICATE list: any specialist conclusion resting on unverified line-items (forensic ratios, exact debt figures, peer multiples) is treated as directional, not quantitative.

8. What I don't know (top 3)

  1. The 12 Boon co-defendants. If any are current or former THG directors, doctors, or suppliers, the "Boon is a family problem, not a hospital problem" thesis collapses and fair value drops 20–30%. Resolved by: AG indictment document or Thai-language press (Thairath/Khaosod) coverage of the charge sheet.

  2. The FY25 related-party note and contingent-liabilities schedule. Whether THG's balance sheet still contains receivables, guarantees, or investments traceable to Boon's Laos/Vietnam phantom projects. Resolved by: FY25 audited 56-1 filing (Form 56-1 currently not accessible).

  3. The post-RAM operational plan and acting CEO. 15 months post-takeover, no named CEO and no margin guidance in the public record. Resolved by: Q4-25/Q1-26 MD&A, opportunity-day deck, or a credible business-press interview with RAM-side management.

9. Triggers to re-rate

Buy more if (move to 0.5% from 0%): - Daily close above ฿8.65 on >5M volume (RAM placement line reclaimed) AND a quarterly print with net margin >3%; OR - Boon case reaches a formal settlement / verdict that explicitly ring-fences THG operating entities; OR - FY25 audited 56-1 confirms no material related-party exposure to Boon-era SPVs

Scale to 1.0% if any TWO of the above trigger together with a named, credentialed CEO and RAM operational margin guidance.

Sell down / stay out if: - Weekly close below ฿7.00 (current support broken — Trader's bearish confirmation) - Any quarter shows return to operating loss - Fresh dilution announcement - Any of the 12 co-defendants is identified as a current/former THG insider

Get out completely if: - Weekly close below ฿4.50 (structural break, orphan→zombie) - AMLO or court order touches THG operating-entity assets - SEC reopens the RAM whitewash waiver - Auditor resignation or qualified opinion appears

Final word to the family: This is a fairly-priced fallen darling with a credible new owner, a fugitive founder, a board we can't fully see, and statements we can't fully verify. The math says ฿7.30 ≈ fair value ฿7.80. We don't pay 0.5% of the book to earn 6% over two years on a stock where the 90% confidence interval is ฿3 to ฿16. We pass, we watch, and we let the chart or the courts tell us when the asymmetry shows up. It is not here today.