Financial Shenanigans

Financial Shenanigans — Tencent Music Entertainment Group (TME)

Tencent Music's reported numbers are broadly faithful to economic reality, but two specific mechanics deserve underwriting attention: the FY2025 IFRS operating profit is inflated by a single non-recurring RMB 2,373m gain on the deemed disposal of the UMG associate stake, and the FY2024 operating-cash-flow surge was structurally carried by a +RMB 1,845m accounts-payable swing that reversed in FY2025. Outside those two items, multi-year disclosure is rich, the auditor's critical-audit-matter language reads conservatively, the balance sheet is net-cash, and DSO has actually improved into FY2025. The principal residual risks are structural — Tencent dependence and a low (sub-1%) accounts-receivable allowance that was rebuilt slowly after a FY2024 release — not active distortion.

1. Forensic verdict — Watch (38/100)

Source: derived from this analysis of company filings [1], [2].

Source: derived from reported financials, FY2023–FY2025 20-Fs [2], [3].

Top two concerns. First, FY2025 IFRS operating profit grew 53.4% to RMB 13,364m but RMB 2,373m of that — roughly 18% of the line — came from a one-time gain on the deemed disposal of the company's stake in an associate (UMG), with the underlying ex-one-off operating profit growing closer to the mid-20s percent [1]. Second, FY2024 reported CFO of RMB 10,275m was lifted by a +RMB 1,845m accounts-payable swing that reversed (-RMB 783m) the following year, so the FY2023→FY2024 CFO jump was not a step-change in cash earnings [2], [3].

Cleanest offsets. The auditor's critical-audit-matter language has been conservative throughout — every audit cycle from FY2021 to FY2025 carried a goodwill-impairment CAM with explicit DCF assumptions, and FY2021–FY2023 carried a second CAM specifically on virtual-gift revenue recognition [4]. Management's own non-IFRS adjusted profit transparently strips the UMG gain via the (Gains)/losses-from-investments line, so investors get a clean ex-one-off picture in every release [5]. DSO actually improved into FY2025 (receivables grew 11.3%, revenue grew 15.8%) [3], and the balance sheet is net-cash with US$300m of 2025 senior notes already repaid [6].

What would move the grade. Downgrade to Elevated if FY2026 working-capital movements again add RMB 1bn+ to CFO and reverse, if accounts-receivable from Tencent Group (RMB 2,249m at year-end 2025) starts ageing, or if the Ximalaya acquisition closes and produces a fresh purchase-accounting reserve build that flatters core trends [7]. Upgrade to Clean if the FY2026 accounts-receivable ECL allowance rebuilds toward the 2.5%-of-gross-AR range that prevailed at end-2023 [8].

2. The headline distortion — the UMG deemed-disposal gain

The single largest forensic adjustment of FY2025 is also the most cleanly disclosed: in Q1 2025 TME recognized a RMB 2,373m gain on the deemed disposal of its associate stake in the UMG-related vehicle, and the entire amount flows through the "Other gains, net" line in operating profit [1]. The associate balance dropped from RMB 4,669m to RMB 1,659m on the year, with a RMB 4,506m "deemed disposal" line flowing through the rollforward, confirming the gain mechanism — the carrying value of the equity-method investment was replaced by a higher fair-value mark when the position was reclassified [9].

Source: FY2024 20-F consolidated income statement (FY2023/FY2024 reported op profit) [10]; FY2025 20-F MD&A and Note 7 Other gains, net [1]. FY2025 ex-UMG is derived from reported FY2025 operating profit less the disclosed RMB 2,373m gain.

Three observations matter. First, management does not hide this: the MD&A explicitly attributes the increase in "Other gains, net" from RMB 165m to RMB 2,632m primarily to "the gain of RMB 2,373 million on deemed disposal of an associate," and the non-IFRS reconciliation strips it out via the (Gains)/losses-from-investments line, taking adjusted profit from RMB 11,353m IFRS to RMB 9,924m adjusted [1], [5]. Second, IFRS net profit growth of +60% is the prominent headline number in marketing material but it overstates the cash-earnings step-up — adjusted profit grew +22%, which is the more honest underwriting anchor. Third, the gain is non-cash: the FY2025 cash-flow reconciliation properly adds back RMB 2,381m of "Net (gains)/losses in relation to equity investments," so reported CFO (RMB 10,231m) is not inflated by the UMG mark — but headline IFRS earnings are [3].

3. The CFO mechanism — FY2024 was a working-capital lifeline

Operating cash flow stepped up by RMB 2.9bn from FY2023 to FY2024, then went sideways into FY2025. The cash-flow reconciliation makes the mechanism explicit: accounts-payable contributed +RMB 1,845m to FY2024 CFO and then reversed by -RMB 783m in FY2025 [2], [3]. That single line accounts for the bulk of the FY2024 "operating leverage" headline.

Source: FY2024 20-F Note 29 cash-flow reconciliation [2]; FY2025 20-F Note 29 [3].

Three forensic implications. First, FY2024's apparent +44% CFO growth was about 60% structural and 40% working-capital timing — a real but smaller business-level improvement. Second, FY2025's flat CFO is better than it looks, because it absorbed the AP reversal and still came in at RMB 10,812m on the back of higher operating profit and a +RMB 538m other-operating-liability tailwind [3]. Third, three-year cash conversion remains strong: cumulative CFO of RMB 27,843m vs cumulative net income of RMB 22,620m gives a 1.23× CFO/NI ratio — the conversion story holds even after smoothing the AP signal.

Source: FY2025 20-F consolidated income statement [11] and Note 29 cash-flow information [3]; free cash flow derived as CFO − capex [3].

4. Receivables and the reserve cycle

Accounts receivable have grown alongside revenue but the loss allowance for expected credit losses dropped from RMB 77m at year-end 2023 to RMB 22m at year-end 2024 before only partially rebuilding to RMB 28m at year-end 2025. The drop was driven by RMB 72m of write-offs against a RMB 17m provision — meaning the net charge through the income statement implicitly absorbed RMB 55m of receivables that had been previously reserved [8], [12]. Coverage collapsed from 2.57% of gross AR to 0.62% over one year and ticked back to only 0.71% in FY2025 — well below FY2023 levels — while the underlying receivable book grew by RMB 936m over the same window.

Source: FY2024 20-F Note 20 Accounts receivable [8]; FY2025 20-F Note 20 [12].

Two points of context discipline the finding before it can be called active reserve management. First, the FY2024 write-offs were largely against social-entertainment-era receivables (the segment whose revenue collapsed from RMB 15,856m in FY2022 to RMB 6,659m in FY2024), and the over-six-month ageing bucket actually fell from RMB 299m at end-2023 to RMB 211m at end-2024 to RMB 118m at end-2025 — supporting the view that bad receivables were genuinely cleaned out [10], [12]. Second, the largest single AR exposure — RMB 2,249m owed by Tencent Group at end-2025 — is functionally credit-risk-free [7]. Net read: yellow flag, not red, but the reserve is sitting near multi-year lows on a growing book and should rebuild if FY2026 revenue mix shifts again.

5. Related-party concentration with Tencent

Tencent Group is both the controlling shareholder and the dominant counterparty: at year-end 2025 RMB 2,249m of TME's RMB 3,931m gross accounts receivable book was owed by Tencent Group, with another RMB 60m owed by Tencent associates — roughly 59% of gross receivables sit inside the parent ecosystem [7]. The income-statement footprint moved sharply in FY2025: revenue from online music services to Tencent Group itself jumped from RMB 188m (FY2024) to RMB 324m (FY2025), revenue to Tencent associates rose from RMB 365m to RMB 501m, and service cost to associates of Tencent Group more than doubled from RMB 592m to RMB 1,399m [7].

Source: FY2025 20-F Note 32 Related party transactions [7]; FY2024 20-F Item 7.B Related Party Transactions [13].

The disclosure language reads as standard for a controlled subsidiary — "these related party transactions were conducted at prices and terms as agreed by the respective parties involved" — without a separately disclosed arm's-length benchmark [7]. This is a yellow flag on structure rather than a red flag on conduct: there is no evidence of round-tripping or circular revenue, the master Business Cooperation Agreement is publicly described, and the Q4 2025 Tencent-related revenue (≈RMB 900m if FY2025 Tencent + Tencent-associates revenue is annualized) is roughly 2.7% of group revenue — small enough that the conduit risk is balance-sheet-driven (the RMB 2.2bn receivable) rather than income-statement-driven.

6. Capitalization, soft assets, and goodwill discipline

The soft-asset story is split. Property, plant and equipment more than doubled — from RMB 721m at end-2022 to RMB 1,488m at end-2025 — driven by FY2021's RMB 2.7bn land-use-right purchase for the Hangzhou headquarters and ongoing construction-in-progress additions [14]. Intangibles built more aggressively too: copyrights additions of RMB 678m hit the book in FY2024 alone, taking intangibles from RMB 2,032m to RMB 2,049m on net but adding RMB 824m to gross cost [15]. Capex of RMB 1,188m in FY2025 ran slightly below D&A of RMB 1,375m, which is consistent with a maintenance-mode asset base rather than aggressive capitalization of opex.

Source: FY2024 20-F Note 29 cash-flow reconciliation [2]; FY2025 20-F Note 29 [3]; FY2025 20-F Note 14 PPE [14].

Goodwill is the larger forensic question, and the answer is defendable but watch the assumptions. The carrying value drifted from RMB 19,493m at end-2022 to RMB 20,521m at end-2025, with no impairment ever recognized — despite social-entertainment revenue collapsing from RMB 15,856m to RMB 6,659m over the same window [10], [16]. The auditor flagged this each year as a critical audit matter. Management's CGU recoverability assumptions disclosed in FY2025 — five-year revenue growth not more than 7% and a 15% pre-tax discount rate — are less aggressive than the FY2024 disclosure (9% and 16%), and combined with a 3% terminal growth rate and the offsetting strength of online-music subscription revenue, the no-impairment conclusion is defensible [16]. The pre-tax discount rate did decrease in FY2025, which adds modeled headroom — worth flagging but not red-flagging.

7. The disappearing virtual-gift CAM

A material disclosure change deserves naming. Through FY2023, the PwC audit opinion carried two critical audit matters: goodwill impairment and "revenue recognition from social entertainment services in relation to sales of durable virtual gifts" — the latter speaking to how long virtual gifts are amortized over (a ≤6-month lifespan estimate) and the resulting timing of revenue recognition. In FY2024 the second CAM was removed, leaving only goodwill [4], and FY2025 keeps that narrowed scope.

The natural rationale is that as social-entertainment revenue collapsed to RMB 6,659m by FY2024 (24% of total revenue, down from 56% in FY2022), the residual audit risk in virtual-gift recognition became immaterial. That reading is plausible but it is also the precise moment when revenue-recognition policy disclosure becomes harder to track for an outside reader — particularly given the FY2025 note language that adds a qualifier on subscription revenue: "the majority of revenue from online music services are recognized over time," with no further breakdown of how much is point-in-time [17]. Yellow flag: the disclosure narrowing coincides with the segment shift, and investors lose a year-on-year comparable CAM.

8. Deferred-revenue smoothing — present but ordinary

Deferred-revenue (contract-liability) balances grew from RMB 3,002m at end-2023 to RMB 3,275m at end-2024, with carried-forward releases of RMB 1,834m (FY2022), RMB 2,170m (FY2023), RMB 2,854m (FY2024) and RMB 3,096m (FY2025) [18], [19]. Releases moving faster than the balance would mark active acceleration of subscription revenue, but releases of RMB 3,096m against a starting balance of RMB 3,275m and a continued build of subscription paying users from 100.9m to 125.1m over the same window read as ordinary subscription growth rather than recognition pull-forward [20]. The advances-from-customers line did jump from RMB 88m to RMB 335m in FY2024 then settle at RMB 208m in FY2025 — a smaller swing that aligns with subscription seasonality rather than smoothing.

9. Cash-flow classification is clean

Three "clean tests" deserve naming because they are not always clean in this sector:

• CF1 — Financing inflows masquerading as operating. The US$800m senior notes issued in September 2020 sit cleanly in financing, the US$300m 2025 tranche was repaid in financing in Q3 2025, and the surviving US$500m 2030 tranche (RMB 3,497m carrying value) has never been confused with operating flows in any release [6]. No evidence of receivables securitization, factoring, or supplier-finance programs in the FY2025 notes.
• CF3 — Acquisition-driven CFO inflation. The FY2025 RMB 1.19bn acquisition of a music-content company added RMB 151m cash acquired, RMB 1,131m intangibles, RMB 806m goodwill — and an explicit disclosure that "the revenue and the results contributed by the acquiree to the Group subsequent to the acquisition were insignificant" [21]. The deemed disposal gain is properly stripped from CFO via the equity-investments add-back line.
• CF2 — Operating outflows hidden as investing. Capex (PPE + intangibles + land-use) of roughly RMB 1.19bn in FY2025 is materially in line with D&A of RMB 1,375m — no evidence of opex being capitalized away from cost-of-revenue. Investing outflow surged because of term-deposit placements (RMB 24bn placed, RMB 18bn matured in FY2024), not because content spend was reclassified out of P&L [22].

10. Non-IFRS hygiene

Non-IFRS adjusted profit for FY2025 of RMB 9,924m is RMB 1,429m below IFRS profit of RMB 11,353m — an unusual sign that the IFRS line is higher than adjusted because the (Gains)/losses-from-investments adjustment of RMB (2,285m) removes the UMG gain. That is exactly what one wants the non-IFRS measure to do, and it gives investors the cleanest single comparison anchor [5]. The adjustment categories — amortization of acquisition-related intangibles, share-based compensation, gains/losses on investments, fair value change on puttable shares, and tax effects — have been structurally stable since FY2022, with no scope creep or new "one-time" carve-outs added year-on-year.

Source: FY2024 20-F Non-IFRS Financial Measure [23]; FY2025 20-F Non-IFRS Financial Measure [5].

The only recurring-cost concern is share-based compensation, which has been adjusted out at RMB 834m (FY22), RMB 736m (FY23), RMB 681m (FY24), RMB 680m (FY25) — between 1.9% and 2.4% of revenue across the four years and clearly recurring [5]. A PM should think about adjusted profit as approximately RMB 680m / year overstated by recurring SBC, which is a standard tech-industry caveat, not a TME-specific shenanigan.

11. The breeding ground — Tencent control, dual class, VIE

The structural conditions raise — but do not amplify into red — the accounting risk:

• Controlling shareholder. Tencent Holdings is described in every filing as the principal owner; the FY2024 20-F names Tencent's Chief Strategy Officer James Mitchell as a director, and the board composition document flags the Tencent affiliation as a non-independent designation [13], [24]. The dual-class voting structure entrenches that control.
• VIE structure. Cayman-incorporated holding entity over PRC VIEs with contractual arrangements — sector standard for U.S.-listed China tech but a structural risk channel that compounds the related-party concern [25].
• Auditor. PwC is the named PCAOB-registered auditor (PCAOB ID 1424) in FY2025, with an unqualified opinion and one CAM (goodwill) [26]. No resignation, no material-weakness disclosure, no scope-of-audit caveats.
• Compensation. Share-incentive plans were unified into a 2024 Share Incentive Plan with the prior plans rolled in; SBC at ~2% of revenue is reasonable and the FY2024 disclosure adopted an insider-trading policy (a material governance upgrade vs the FY2023 filing's "Not applicable" line) [27].
• Ximalaya merger pending. The June 2025 Merger Agreement to acquire Ximalaya for US$1.26bn cash plus up to 5.57% of TME's Class A shares is the next forensic battleground — purchase-accounting reserves, contingent consideration treatment, goodwill allocation, and any subsequent step-up gains will need close reading once the deal closes [19].

Net read: the breeding ground is structural rather than predatory. Tencent dependence is the load-bearing risk; weakness of independent oversight is structural; auditor and disclosure cadence are sector-leading.

12. The 13-category shenanigans scorecard

Sources for the scorecard rows: FY2025 20-F Item 5.A Operating Results & Non-IFRS reconciliation [1], [5]; FY2024 20-F Note 29 and Note 20 [2], [8]; FY2025 20-F Notes 7, 16, 17, 20, 27, 32 [7], [12], [16].

13. What to underwrite next

1. FY2026 Q1–Q4 working-capital movements. If accounts payable adds another RMB 1bn+ to CFO and reverses, the "lifeline" pattern is confirmed. Watch the AP and other-operating-liability lines in each quarterly press release [28].
2. Ximalaya close mechanics. When the US$1.26bn + 5.57%-equity deal closes, scrutinize: (i) the purchase-price allocation between goodwill, intangibles, and acquired working capital; (ii) any step-up gain on previously held equity interests; (iii) earn-out / contingent consideration measurement; (iv) whether merger-related charges get excluded from non-IFRS as "one-time" while the integration drags on [19].
3. AR allowance rebuild. A walk from 0.71% to ~2% coverage in FY2026 would resolve the EM5 yellow flag; a flat or further reduction would harden it [12].
4. Tencent-related revenue / AR trajectory. A continued ramp in Tencent-direct online-music revenue (RMB 188m→324m in FY2025) into FY2026 — particularly if not matched by paying-user growth — warrants a follow-up question on counter-party terms and pricing benchmarks [7].
5. Goodwill CGU assumptions. The FY2025 disclosed pre-tax discount rate dropped from 16% to 15%; another step-down without underlying improvement in CGU performance would be the kind of assumption easing that should be challenged in the next CAM read [16].

Bottom line for the position file. Accounting risk at Tencent Music is a valuation-haircut item, not a thesis-breaker. The structural Tencent dependence and the FY2024 CFO lifeline justify (a) anchoring earnings on non-IFRS adjusted profit, not IFRS net income, and (b) using a three-year average CFO (RMB 9.28bn) rather than any single year as the cash-flow input for valuation. With the UMG gain stripped, FY2025 adjusted profit grew at the mid-20s percent and three-year CFO/NI is 1.23× — strong but not extraordinary. The receivable-reserve and related-party concentration are real but bounded by transparent disclosure. A PM should not pass on TME for forensic reasons, but should size the position with the understanding that one specific quarter (FY2024 Q4) and one specific line (FY2025 other gains) are doing more of the headline-growth work than they look like at first glance.