Financial Shenanigans

Figures converted from KRW at historical period-end FX rates — see data/company.json.fx_rates for the rate table. Ratios, margins, and multiples are unitless and unchanged.

The Forensic Verdict

SKC's accounting is not the source of the problem — the business is. But the way the numbers are presented makes the underlying damage harder to underwrite than it should be. The continuing-operations P&L has been sub-zero for three straight years, yet pretax losses have run roughly 2.4x to 3.0x larger than reported operating losses (FY25 OP of -$211M versus pretax of -$620M), with the missing -$410M buried in below-the-line items the English filings never break out. The full audited cash-flow statement is filed only on DART in Korean, so investors outside the home market cannot directly verify CFO. Repeated divestitures — PET film (2022), CMP Pad/FCCL/Blank Mask/CMP Slurry (2024-25), Fine Ceramics (2024) — re-baseline "continuing operations" each year, and Q4 2025 layered visible big-bath signals in EV battery materials and Chemical (write-down of "manufacturing-equipment optimization") just as a rights offering was being announced. We grade the file Elevated, not High — there is no restatement, auditor change, or material-weakness disclosure, the audit committee is fully independent and active, and balance-sheet hygiene metrics like DSO are clean. The single data point that would change the grade is a clean, English-language consolidated cash-flow statement showing that 2025's $443M cash build came from genuine working-capital release plus the disclosed $616M asset-rebalancing program — not from receivable factoring or supplier-finance lifelines. Until that arrives, treat headline cash and headline operating loss as informationally incomplete.

Grade: Elevated. The accounting risk here is opacity-driven, not manipulation-driven. The English-language filings do not lie; they under-disclose. A reader cannot, from public English data alone, confirm where -$410M of FY25 below-the-line losses came from, why non-controlling interests gained $10M in a year the parent lost $507M, or how $443M of cash was actually built. None of these is a fraud signal in isolation. Together, they require an underwriting discount.

Shenanigans scorecard

Breeding Ground

The breeding-ground signal is concentrated ownership inside a chaebol ecosystem, balanced by a board that looks well-functioning on paper. The amplifier and the dampener largely cancel.

The board structure does what governance is supposed to do — independent majority, fully-independent audit and internal-transactions committees, no opposed votes but two modified votes recorded in 2022, ISO 37301 compliance certification newly obtained. The KCGS named SKC an "Excellent Company" in governance in 2024. Against that, the parent SK Inc. holds roughly 40% of shares, the major operating subsidiaries are co-owned with related parties (Hahn & Co. at SK nexilis, Apollo at Absolics), and detailed director compensation is disclosed only on DART in Korean. None of this is unusual for a Korean chaebol affiliate; it does mean an English-only investor is reading through a smaller window than a domestic counterpart.

The single breeding-ground signal that did not fire — the streak of beating expectations — fails to fire because there is nothing to beat: SKC has reported losses for three straight years. That cuts both ways. The earnings-management incentive to flatter results is muted; the survival incentive to time write-downs and divestitures around capital raises is not.

Earnings Quality

Earnings quality has three tiers: the operating line, the below-the-line gap, and the discontinued-operations cushion. Only the first looks like it reflects underlying economics. The other two require structural skepticism.

The pretax-to-operating gap

The widening wedge between the blue and red bars is the most material forensic signal in the file. From a clean $4M gap in FY22, non-operating items deteriorated to -$142M in FY23, -$262M in FY24, and -$410M in FY25. The English deck does not disaggregate this line. Plausible drivers — equity-method losses on Absolics (US glass-substrate JV, pre-revenue), impairment of investments in associates, finance costs on a leveraged balance sheet, FX losses, and likely write-downs of intangible assets — are individually visible elsewhere in the file (intangibles fell from $1,287M to $915M over FY23-25), but the income-statement attribution is missing.

This pattern is exactly what the playbook calls "hiding losses below the line": OP is the metric management leads with; the much larger pretax loss arrives later and is rarely re-quoted.

The discontinued-operations cushion

In four of the last six years, discontinued operations were materially larger than continuing operations — sometimes a tailwind (FY20, FY21, FY24) and sometimes a drag (FY22, FY23). FY24 is the most consequential: a +$84M discontinued-ops gain softened the headline net loss from -$393M (continuing) to -$309M (reported). In FY25 the company appears to have closed the divestiture cycle, so the cushion disappeared and the full continuing-operations loss flowed through. The optical improvement that any FY26 print could deliver should be measured against continuing-ops earnings only, not against the cushioned net-income line.

A related governance note: the number of consolidated entities collapsed from 42 in FY23 to 22 in FY24. That 20-entity reduction reshapes the comparable group and complicates any like-for-like FY24 vs FY23 read.

Margin and reserve hygiene

Operating margin has stabilized around -16% for two years, but net margin keeps deteriorating (-21.9% → -26.4% → -39.1%). The widening of the net-vs-OP wedge — from 7.5pp in FY23 to 22.5pp in FY25 — is the same below-the-line story restated as ratios.

Cash Flow Quality

This section is partially blocked by disclosure: SKC's full audited cash-flow statement is published only in Korean on DART. The English Sustainability Report and 4Q earnings release deliberately omit it. We can reconstruct directional cash flow from balance-sheet changes and from disclosed corporate actions, but we cannot run the standard CFO/NI test or accruals test cleanly.

The FY25 cash spike — from $274M to $721M, a +$447M build in a year of -$496M net loss — is the question. Management's own 2025-review slide answers most of it: $616M of asset rebalancing (CMP Pad $231M + FCCL $66M + Blank Mask $47M + CMP Slurry $8M = $351M of disposals, plus $266M of exchangeable-bond issuance). Reconciling roughly: -$496M loss + non-cash D&A ~$130M + disposal proceeds $351M + EB $266M + working-capital and FX = ~$275-450M cash build, in the right ballpark for the observed +$447M.

Inventory rebuilt from $160M to $235M in FY25 — that consumes cash, which is a clean signal: cash strength was not built by under-purchasing inventory. Trade receivables for FY25 are not separately disclosed in the 4Q earnings release (the deck shows only aggregate current assets), which is itself a yellow disclosure gap.

The forensic conclusion is: FY25 cash flow looks reasonable but is fundamentally non-recurring. Two large divestitures and an exchangeable bond covered three quarters of the build. Management explicitly described 2026 priorities as "cash flow-centric management" with "CAPEX discipline" and "non-core asset rebalancing option" — language that telegraphs continued reliance on disposal-driven cash and capital-markets access. A rights offering was announced in 1Q26 and is scheduled to list June 8, 2026.

Metric Hygiene

The metrics SKC leads with — segment revenue, segment OP, EBITDA, and the YoY/QoQ deltas in the 4Q deck — are consistent with the underlying statements but are framed to make the quarter look better than the cumulative print does.

The most underweighted metric in the file is the divergence between parent equity and non-controlling interests. Parent equity collapsed from $1,681M in FY21 to $574M in FY25 (a -66% draw in USD terms, -58% in home currency); NCI rose from $236M to $824M over the same window. That means co-investors at the subsidiary level — Hahn & Co at SK nexilis, Apollo and SK Inc. at Absolics — funded growth capex while the listed parent absorbed an outsized share of the operating losses. In FY25, parent net loss was -$507M against group net loss of -$496M, implying NCI booked a +$10M gain in a year the parent burned through $507M. That asymmetry is allowed under K-IFRS when losses occur in entities where the parent funded most of the equity, but it is not reader-friendly: the ratio that matters for SKC's listed-share holders is parent-attributable, which is meaningfully worse than the group line.

Tangible assets ramped $1,563M → $2,248M over FY22-24 (+44% in USD; +67% in home-currency terms) — the copper-foil capacity build in Malaysia and Poland plus Absolics in Georgia. Soft assets (intangibles + goodwill) drifted down $1,287M → $915M (-29% in USD) over FY23-25, suggesting some impairment did flow through, though the magnitude is not separately tabulated in the English filings. The capex/D&A ratio in FY24 was approximately 4.2× depreciation — deep growth-mode spending into a market that is now penalizing copper-foil supply with chronic operating losses. The forensic question for FY26 is whether further intangible/PP&E impairment will be required if Malaysia and Absolics ramps fall behind plan.

What to Underwrite Next

The forensic risk here does not change the direction of the investment thesis — it shapes the position size and the margin of safety required against the operational thesis. Five specific items belong in the next-quarter and next-AR diligence list:

1. The full audited cash-flow statement (DART, in Korean) for FY24 and FY25. The single missing piece. It would let an analyst test CFO/NI, accruals ratio, and working-capital contribution directly. Without it, one is reasoning from balance-sheet deltas. What changes the grade: a clean CFO ≥ -$140M excluding asset disposals, with working-capital contribution under $70M, would downgrade the file to "Watch."

2. Disaggregation of the $410M FY25 below-the-line losses. Specifically: equity-method losses on Absolics, finance costs, FX, impairment of investments in associates, and any goodwill/intangible impairment. Each is recoverable from the K-IFRS notes; they are simply not in the English deck. What changes the grade: if the bulk is finance cost + Absolics equity-method losses (recurring but predictable) rather than impairments (concentrated and lumpy), it is less concerning.

3. The post-rights-offering pro-forma balance sheet. The June 8, 2026 listing materializes the dilution. Watch parent equity, current ratio, retained earnings (currently negative at -$35M), and net debt for the recovery profile. What changes the grade: parent equity restored above $830M and current ratio above 1.0 in 2Q26 print would downgrade the file to "Watch."

4. Quantification of the Q4 2025 "one-time expenses." The 4Q deck describes "Recognition of one-time expenses" in EV battery materials and a "write-down of manufacturing equipment optimization" in Chemical without giving USD magnitudes. The 1Q26 deck already uses the phrase "narrowed operating loss given one-time costs recognition in 4Q25" — i.e. management is preparing the comparison. Watch FY26 OP for "ex-one-time" framing that would not have surfaced in a normal quarter.

5. Related-party transaction volume with SK Group affiliates. SK Inc. is the ~40% controlling shareholder. The Internal Transactions Committee meets 8-12× per year. Aggregate annual transaction volumes with SK affiliates (SK Inc., SK Discovery, SK Chemical, SK enpulse remnants) are filed in Korean only. What changes the grade: a sudden expansion of intra-group transaction volume during the rights-offering window would raise the file toward "High."

The file is Elevated, not High. Move it down to "Watch" when the consolidated cash-flow statement and the FY25 below-the-line breakdown become available in English. Move it up to "High" only if a new auditor finding, a fair-value adjustment dispute on Absolics, or an undisclosed related-party transaction appears.