Skip to content

FORENSIC ACCOUNTANT REPORT — XPG (X Spring Capital PCL)

PRELIMINARY NOTE — THIS IS NOT A FORENSIC ACCOUNTING REPORT, BECAUSE THERE ARE NO ACCOUNTS TO FORENSICALLY ACCOUNT.

The data package contains zero extractable financial statement data. The SET shareholders page and filings page returned HTTP 404 ("ขออภัย ไม่พบข้อมูลที่คุณต้องการ" / Error 404). The SET profile, financial-highlight, and news pages returned only Nuxt JS-shell HTML — no rendered figures. No income statement, balance sheet, cash-flow statement, MD&A, 56-1 reference, auditor name, related-party note, segment breakdown, share-pledge disclosure, or shareholder list is present. The only XPG-specific news item in the package is a Bangkok Post headline from 29 May 2021 ("Investor sells chunk of KTC for XPG venture"), now five years stale relative to the package's 2026-06-23 timeline.

Per the Data Audit's DO-NOT-FABRICATE list, I am forbidden from quoting revenue, net income, EPS, ROE, NIM, AUM, P/E, P/B, BVPS, shares outstanding, shareholder percentages, or computing Beneish/Altman/Sloan — all required inputs are absent.

What follows is therefore not a forensic verdict on XPG's books. It is a structured statement of what cannot be assessed, with the tape-based signals that can be read, and the specific line items a credible workup would require. A forensic accountant who pretends to grade earnings quality off a JS shell is not a forensic accountant — they are a fabulist.

1. Overall accounting quality grade: UNGRADEABLE (effectively F on disclosure-to-this-package)

Scoring on the basis of what the data package allows me to verify, not on XPG's actual books (which I have not seen):

  • Earnings quality: N/A / 10 — no income statement in package.
  • Cash conversion: N/A / 10 — no cash flow statement in package.
  • Balance sheet health: N/A / 10 — no balance sheet in package.
  • Disclosure transparency (this package only): 1 / 10 — two of five SET pages 404'd; three rendered only <head> boilerplate; zero current company-specific news; auditor unknown; shareholder list unknown; no 56-1 referenced. Note this scores the package, not the company — the SET likely hosts these data; they simply did not arrive here.

A real accounting-quality grade for XPG requires the actual filings. They are not in the package.

2. The 10-year cash-vs-earnings picture

Cannot compute. No cash flow from operations figure, no net income figure, no annual series for any year 2016–2025 is present in the package. The 10-year "summary row" provided (total_return=6.99, ann_return=15.12, ann_vol=52.73, sharpe=0.287, max_dd=−90.8) is price-return metadata, not accounting data, and per the Data Audit is internally inconsistent with first=0.4206 → last=0.45 and must not be cited as fact.

What the price tape does tell a forensic reader: - Max drawdown of approximately −90.8% over the 10-year window (audit-flagged but qualitatively plausible for a Thai penny financial that restructured in 2021) is consistent with a regime in which either (a) heavy dilution occurred, (b) book value was destroyed, or (c) both. Any of those routes warrants forensic attention to share-issuance history and impairment patterns — neither of which is in this package. - The current price band of THB 0.43–0.49 (Mar–Jun 2026 tape) sitting at a sub-THB-0.50 floor with no material rallies suggests the market is not pricing accruals optimism; it is pricing distress or stasis. That is a market read, not an accounting read.

3. Working capital diagnosis

Cannot compute. Receivables days, inventory days, payables days, DSO trend, accruals-to-assets — none of the line items required (trade receivables, inventories, trade payables, revenue, COGS) are in the package. For a financial holding company (X Spring Securities + asset management + PE), the relevant working-capital analogues would be: securities receivables/payables, margin loans, client cash payable, repo book — also absent.

4. Revenue quality

Cannot assess. Required disclosures all missing from package: - Revenue concentration by client / fund — not in package. - Related-party transaction note — not in package. - Brokerage commission vs. fee income vs. investment gains split — not in package. - Revenue recognition policy under TFRS 15 / TFRS 9 for the securities and asset-management arms — not in package. - Single-customer concentration disclosures — not in package.

Structural concern carried over from the issuer profile only (config note, not a primary filing): a Ravi-Chanchareonsook-controlled investment-holding company with a private-equity arm, an asset-management arm, and a captive broker is structurally exposed to (i) fair-value gains on Level-3 PE positions flowing through P&L, (ii) intra-group transactions between the broker and the PE / AM arms, and (iii) management-fee revenue from related-party funds. These are the line items a forensic workup must extract from the 56-1 — but the 56-1 is not in this package.

5. Margin trajectory & one-offs

Cannot compute. No gross margin, EBITDA margin, EBIT margin, or "other income" line is present for any year. The 2021 restructuring event referenced in the issuer note would, in a complete dataset, almost certainly show as a discontinuity in margin series and likely a large one-off gain or loss — but that series is not in this package.

6. Capex / depreciation / asset base

Cannot assess. No PP&E, capex, depreciation, or intangibles roll-forward in the package. For a financial-services holding, the more relevant analogue — investment book mark-to-market, goodwill from the 2021 reverse-acquisition / rebranding, and intangibles capitalisation — is also absent. Goodwill from the 2021 transformation (if any) is precisely the kind of asset that warrants impairment scrutiny in a sub-par-trading penny stock. Cannot test this without the balance sheet.

7. Off-balance-sheet & leverage red flags

Cannot assess. Required disclosures all absent: - Debt at parent vs. subsidiary — not in package. - Operating leases / lease liabilities under TFRS 16 — not in package. - Financial guarantees to subsidiaries / related parties — not in package. - Insider share pledges — critical for a Ravi-controlled vehicle trading at THB 0.45 with −90% historical drawdown; would normally appear in the SET shareholders page and 56-1; shareholders page 404'd, 56-1 not in package. This is the single most important missing disclosure for this issuer. - Margin-financing book and repo exposure at the securities subsidiary — not in package.

The Data Audit flagged a Nation Thailand article (23 Jun 2025) about "seven Thai stocks hit floor; forced selling suspected." XPG involvement is not confirmed in the package excerpt. If XPG were one of those seven, that would be direct evidence of pledged-share forced-sale dynamics — but the package does not establish the link, and I will not infer it.

8. Auditor & policy changes

Cannot assess. Auditor name not in package. Auditor history not in package. Accounting policy notes not in package. Restatements not in package. The 2021 restructuring almost certainly involved accounting policy choices (purchase-method vs. reverse-acquisition treatment, fair-value step-up on the acquired investment portfolio, recognition of intangibles) that a forensic reviewer would test against the filed 56-1 — the 56-1 is not in this package.

9. Quantitative red-flag scores

Per the Data Audit DO-NOT-FABRICATE list, and because no balance sheet or income statement line items are in the package:

  • Beneish M-Score: Cannot compute. Requires DSRI, GMI, AQI, SGI, DEPI, SGAI, LVGI, TATA — zero of the eight inputs are available.
  • Altman Z-Score (or Z″ for non-manufacturers): Cannot compute. Requires working capital, retained earnings, EBIT, market cap, total liabilities, total assets, revenue — zero of seven inputs are available. (Note: classical Altman is in any case inappropriate for a financial-services holding; Z″ or a bank-specific distress model would be needed. Both still require the balance sheet.)
  • Sloan Accrual Ratio: Cannot compute. Requires net income and cash flow from operations across two consecutive year-ends — none available.

A persona that produces M, Z, or Sloan numbers from this package is fabricating them.

10. Top 5 specific concerns, ranked

These are concerns about what we cannot see, not findings against the company:

  1. Shareholders page 404 + share-pledge disclosure missing. For a controlled-shareholder, sub-THB-0.50, post-restructuring vehicle, pledged-share / forced-sale risk is the single biggest forensic question. The package answers it with a 404. Fair-value discount until cleared: minimum 15–25% relative to an issuer with verified pledge-free control block.
  2. Filings page 404 + auditor unknown. Cannot identify auditor, audit opinion, key audit matters, or restatement history. For a financial holding with Level-3 PE marks, the auditor's identity and tenure is itself a quality signal. Fair-value discount: 10–15% until auditor name and opinion verified.
  3. No 56-1, no segment financials, no related-party note. The whole forensic risk surface for an investment-holding company lives in the related-party section and the fair-value hierarchy disclosure. Both invisible here. Fair-value discount: 10–20% until related-party flows quantified.
  4. Five-year news-flow blackout. The only XPG-specific English headline in the package is from 29 May 2021. Either coverage genuinely collapsed (consistent with a low-float, low-liquidity holding company drifting off the sell-side radar) or the package's news scrapers failed. Either way, there is no public-market check on management narrative. Fair-value discount: 5–10%.
  5. 2021 restructuring leaves an accounting-policy crater that cannot be inspected. Reverse-acquisition accounting, goodwill recognition, fair-value step-ups on the investment portfolio, and recognition of brand/customer-relationship intangibles all would have happened in 2021. Five years on, those goodwill/intangible balances are the prime impairment candidates given the −90% historical drawdown. Cannot test from this package. Fair-value discount: cannot quantify; this is a "do not invest until inspected" item.

11. What's clean and not a problem (be balanced)

In fairness, and to avoid the forensic accountant's occupational vice of treating absence of data as evidence of fraud:

  • The price tape is internally consistent on every shown OHLC row (low ≤ open/close ≤ high). No printing irregularities.
  • Zero-volume "stale close" days on 2026-05-01, 05-04, 06-01, 06-03 align with Thai public holidays — this is a yfinance feed convention, not evidence of trading suspension or a halt.
  • The penny-price range is stable (THB 0.43–0.49 across 60 sessions) with no pump-and-dump pattern visible in the recent window — no late-day ramps, no end-of-quarter spikes, no obvious painting-the-tape.
  • The SET pages that did render were the real SET pages (correct title tags, real Nuxt SSR scaffolding, correct OG/Twitter metadata) — the issuer is genuinely SET-listed; this is not a phantom ticker.
  • The 404s on shareholders/filings are most likely a SET URL-routing or scraper-path issue, not evidence that XPG has been delisted or struck off — the profile/financial/news pages exist with normal <title> tags.

Bottom line for the CIO

I cannot grade XPG's accounting quality from this package. Anyone who claims to has either separately sourced the 56-1 and audited financials (and should say so) or is fabricating. The correct next step is not investment, not divestment, but data acquisition: pull the 56-1 for FY2022–FY2025, the latest annual report, the auditor's opinion, the shareholders' meeting circulars (for share-pledge disclosure), and the SETSMART financial-statement extract. Then this report can be written.

Word count: ~1,640.