Financial Shenanigans

Financial Shenanigans — Wise plc (WISE)

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

1. The Forensic Verdict

Risk grade: Elevated (45/100). Wise's reported numbers reconcile, the auditor (PwC) has not flagged a control issue, receivables actually shrank in FY2025 while revenue grew, and the headline IFRS gross-up of customer balances on cash-flow and balance-sheet lines is a structural feature of any e-money payment institution rather than aggressive accounting. What pushes the grade above "Watch" is a different problem: a CEO with a 49.3% voting bloc who was personally fined by both HMRC and the FCA, a dual-class structure that was just extended ten years to 2036 in a vote that ISS and Glass Lewis say they were briefly misled about, three control-environment leadership changes (CFO, Chief Compliance Officer, Head of Internal Audit) inside a twelve-month window, and three live regulator settlements in the US, UK and ADGM — all wrapped around a "Underlying PBT" framework, introduced for FY2024 reporting, that excludes the most volatile component of earnings (interest income above the first 1% yield). The single data point most likely to change the grade is the PwC audit opinion language for FY2026, the first year the new compliance and finance leadership team owns the books and the first year of the dual-listing transition.

The CFO/NI ratio of 10.8x is mechanical: cash held on behalf of customers ($22.1bn) sits inside trade payables and is matched by safeguarded cash and short-term investments, so growth in customer balances flows through CFO. The economically-meaningful cash measure is management's own "Underlying free cash flow" of $431m in FY2025, which sits at 118% of underlying PBT — clean, not stretched. Investors who cite the headline $5.82bn CFO are quoting a number that is not theirs.

Shenanigans scorecard

Three categories drive the elevated score: metric architecture (#12), customer-balance presentation (#13) and the related cash-flow gross-up (#8). The earnings-quality column is mostly clean — the company is not playing games with revenue recognition, accruals, capitalisation or impairments.

2. Breeding Ground

The structural setup tilts toward management discretion, not external challenge. Käärmann's economic stake of roughly 18% converts to 49.3% voting via Class B 9-vote shares; Notorious OÜ (his investment vehicle) holds another 13.1% of voting rights. The dual-class sunset has just been extended from July 2026 to 2036 inside a bundled vote on the move of the primary listing to the US, a bundling that ISS and Glass Lewis publicly admitted they "initially failed to identify" and that co-founder Taavet Hinrikus opposed. Käärmann's voting power is contractually capped at one vote below 50% while CEO and one vote below 35% if he steps down — meaningful only because without the cap he would otherwise control more than 50%.

The redeeming factor is that Käärmann himself takes no variable pay — he draws a $255k salary and has no bonus, no LTIP, no PSP. That removes the most common compensation lever for accounting aggression. The next most senior executives, including new CFO Thomassin, do have LTIPs at 400% of salary, weighted 40% TSR / 20% cross-currency volume / 20% Underlying PBT margin / 20% NPS. Volume growth as an incentive line is a flag worth monitoring because it can be hit by reducing take-rate — already a stated strategy in the FY26 transcripts ("Mission Zero" pricing) — without earnings discipline. The Underlying PBT margin component has a 10% threshold floor against a 13-16% guidance range, which gives the Remuneration Committee meaningful discretion to vest below the publicly-communicated band.

PwC has been auditor since the 2016 SE-domicile period; senior partner Mark Jordan rotates after FY2026 and there is no audit tender planned for FY26 — investors should treat the FY26 audit opinion (the partner's last) as the cleanest backward-looking signal available before the move to US GAAP and a new audit relationship.

3. Earnings Quality

Reported earnings look earned, with one architectural caveat. Revenue, gross profit, operating profit and net income tie out cleanly across years, the gross margin trend (FY2022 66% → FY2025 79%) is consistent with management's stated take-rate reduction being more than offset by mix shift to high-margin Card and Other revenue lines, and there is no contract-asset, capitalised-cost or other-asset balance that has decoupled from revenue. The architectural caveat is that the largest contribution to reported pretax income is now interest income on customer balances ($769m FY25, +71% YoY despite falling rates) — booked separately on the income statement, but pushed below the "Underlying PBT" measure management directs investors toward.

The FY2024 single-year mismatch — receivables grew 77% while revenue grew 47% — is worth noting but is not a revenue-recognition red flag. Wise's "receivables" line is dominated by amounts due from card schemes and partner banks in transit, which scale with card spend; FY2024 was the first year of material scale for the Wise card business. The FY2025 reversal — receivables down 22% on revenue up 17% — is the test that matters and it passed cleanly.

The FY2024 step-change in operating margin (16.3% → 35.5%) is the single most important earnings event in this dataset and deserves attention. It is not a revenue-recognition or capitalisation event — it is the first full year in which interest income on customer balances ran at materially elevated rates (the Bank of England base rate stayed at 5.25% for most of the year), and management chose to retain the entire excess yield as profit rather than pass it through to customers. Whether that margin is sustainable is the central business question; whether it was earned is not in dispute.

The framework caps the interest-income contribution to "Underlying income" at the first 1% yield. In FY2025 that meant $575m of interest income was excluded from the lead earnings metric, alongside $209m of customer rebates that fund the high-yield customer accounts ("Wise Interest" / "Wise Assets" benefits). The reconciliation is fully disclosed and consistent across FY24 and FY25, but it does mechanically smooth the underlying margin against the rate cycle — a feature management will lean on as rates fall in FY26 and again has the practical effect of preserving the 13–16% Underlying PBT margin guidance band even as reported PBT margin moves with the rate environment.

Capitalisation, reserves and impairment behavior are clean. Capex was $44.7m in FY2025 (2.1% of revenue), up from $13.4m the prior year because of the Brussels HQ build-out and infrastructure investment — not customer-acquisition or capitalised-software inflation. Intangibles on the balance sheet are $5.2m (down from $8.2m), so there is no goodwill-amortisation or intangibles-build to manage. There were no restructuring charges, no impairment events, no big-bath behavior around the CFO transition, and no inventory or accrual line worth flagging.

4. Cash Flow Quality

The headline cash flow statement is uninformative without the customer-balance overlay. Reported operating cash flow of $5.82bn in FY2025 — 10.8x net income — is overwhelmingly the IFRS-mandated gross-up of customer balance growth flowing through working-capital lines. This is not a forensic distortion in the traditional sense (it is the only correct presentation for a regulated e-money institution), but it makes the headline CFO/NI and FCF/NI ratios the wrong tool for assessing this business. Management's disclosed Underlying Free Cash Flow of $431m (FY2025) is the right benchmark and is converting at 118% of Underlying PBT — fundamentally clean.

The right way to read the chart above is that the $5bn-plus gap between net income and CFO each year is almost entirely the change in customer balances (trade payables grew from $17.6bn to $22.8bn in FY25 alone, a $4.8bn move). Wise's economically-controllable cash conversion is a different — and reasonable — story.

The $18bn-plus delta between reported "free cash flow" and the underlying measure is a real disclosure risk if a downstream investor or screener pulls the headline number from the cash-flow statement and treats it as cash generation. It is not an accounting choice Wise can change — IFRS dictates it — but it is a metric that needs the management bridge to be useful.

There are no working-capital lifelines, factoring, supplier finance or receivable sales disclosed. There is, however, one financial-engineering layer worth flagging: in FY2025 Wise put in place a $673m "Safeguarding Insurance" policy that allows the company to use a portion of safeguarded customer funds for operational liquidity (sized to FCA-permitted limits). This is a regulatory innovation, not an accounting one — but it does change the practical liquidity dynamics, and is the new mechanism that should be monitored alongside the $194m revolving credit facility (now unsecured following the November 2024 BBB ratings from S&P and Fitch).

5. Metric Hygiene

This is the single biggest forensic concern in the file. Wise's metric architecture changed in FY2024 — the lead earnings metric switched from "Adjusted EBITDA" (with a stated target of "20% or higher" margin) to "Underlying PBT" (with a stated guidance of 13-16% margin). The Underlying PBT framework excludes the most volatile component of earnings — interest income above the first 1% yield — by construction. The framework is internally consistent, fully reconciled and stable across FY24 and FY25, but it is unambiguously a presentation choice that softens the headline-margin versus target comparison and smooths reported earnings against the interest-rate cycle. It is also the metric that drives executive compensation.

The change of lead metric from Adjusted EBITDA to Underlying PBT was justified at the FY24 results call as "the metric that best reflects how we run the business." That is a defensible position — Underlying PBT does include depreciation and stock-based compensation, which makes it stricter than the prior EBITDA frame in those respects. But the simultaneous decision to cap interest-income recognition at the first 1% yield is the part that directly affects comparability: every basis point of base rate above that floor is excluded from the management metric and parked in the reconciliation. A 1.0% policy rate environment would put 100% of interest income in the underlying metric; a 5.0% rate environment puts roughly 25% of it in the underlying metric. The metric is therefore most generous, relative to the alternative of including all interest income, exactly when interest income is highest — which is the period now ending.

Three top-line numbers, all roughly $2.0-2.7bn, all internally consistent, all used in different parts of the FY2025 reporting package. Investors who anchor on "Income" risk overstating the size of the operating business by 36% relative to IFRS revenue; investors who anchor on Underlying income risk understating the contribution of the rate environment to FY24-FY25 reported earnings.

6. What to Underwrite Next

The accounting risk does not appear large enough to be a thesis breaker, but it is large enough to require a margin of safety — somewhere between a position-sizing limiter and a small valuation haircut on the rate-driven component of earnings. The fundamental shenanigan signature is absent: revenue is earned, receivables are clean, capex and intangibles are tiny, no factoring, no acquisitions to mask, no big-bath behavior, auditor is mid-tenure with no qualification language. The risk that does remain is concentrated in three areas: governance (one shareholder controls the vote, and that shareholder has been personally fined by two regulators), metric definition (an Underlying PBT framework that systematically excludes the cyclical component of earnings exactly when it is highest), and compliance (three live regulator settlements with one independent monitor in place, in a business whose fundamental input is regulated cross-border money movement).

Track these five things in the next two reporting cycles:

1. PwC FY2026 audit opinion language. Partner Jordan rotates after FY26 and the company is preparing to switch to US GAAP for the dual-listing. Any emphasis-of-matter, key-audit-matter elaboration, or change to the customer-balance critical-judgement disclosure is the highest-signal observation available before the new auditor starts.
2. The $45m dual-listing transition costs. If these are formally moved into a "non-underlying" reconciling item in FY26 reporting — separate from the regular Underlying PBT framework — the metric architecture has loosened. If they are absorbed inside underlying PBT, the discipline holds.
3. MMET independent monitor reporting. The July 2025 consent order with the US Multi-State MSB Examination Taskforce includes a multi-year independent monitor; any extension, escalation or finding from that monitor is the most concrete near-term compliance read.
4. Underlying PBT margin in a falling-rate environment. H1 FY2026 already showed the framework holding (16.3% margin vs 22% prior period) but excluding the $45m one-offs. The FY26 full-year print is the first full-cycle test of whether the 1%-yield framework actually delivers the 13-16% guidance band when rates fall, or whether the framework produces a margin print well below the band that requires a further definitional adjustment.
5. Replacement of Chief Compliance Officer. Steyn resigned 31 December 2024 with an interim in place. A permanent CCO appointment with a clear remediation mandate is the cleanest signal that the FY25 risk-factor language ("strategically strengthen our sanctions and AML/TF controls, remediate all known control gaps") is being executed.

Three signals would upgrade the grade toward Watch (sub-30): a clean PwC FY26 opinion, MMET monitor exit, and a permanent CCO appointment. Three signals would downgrade it toward High (above 60): a restatement of Underlying PBT or the income-framework definition, a regulator finding that broadens beyond AML into product economics, or any indication that the $45m dual-listing costs become a recurring "non-underlying" line. The forensic finding here is not that the numbers are wrong — they are not — but that the architecture around the numbers has moved in the company's favor at the same time as the governance structure has been entrenched for another decade. That combination warrants underwriting, not abstention.