Financial Shenanigans

Financial Shenanigans — Synchrony Financial (SYF)

Forensic verdict: WATCH (score 37/100). Synchrony's reported numbers are, on the whole, a faithful representation of economic reality — the audit is clean, internal control over financial reporting is effective, cash flow is honestly presented, and there is no restatement, regulatory accounting action, or auditor concern [10]. But two things keep this off "Clean." First, the headline earnings growth of the last two years is not core operating growth: FY2024's reported +56% jump in net earnings was almost entirely a one-time disposal gain, and FY2025's growth leaned on a $1.5 billion drop in credit-loss provision (a reserve release), not on revenue [1][2]. Second, the single largest estimate on the balance sheet — the allowance for credit losses — was cut to its lowest coverage since CECL adoption in the same year management changed the loss-forecasting model, and that estimate is the auditor's only critical audit matter [3][7][9].

Sources: FY2025 10-K, Summary Highlights & Statistical Data [1][3]; FY2024 10-K, Highlights [2]; CFO/NI and net-earnings comparison derived from reported financials.

Synchrony is a monoline U.S. consumer lender — private-label and co-branded credit cards funded mostly by direct deposits. For a lender, the forensic risk does not live in revenue cut-off or channel stuffing; it lives in reserve adequacy, the timing of provision, and the per-share optics of a relentless buyback. That is where this memo spends its space. Categories that do not apply (bogus/related-party revenue, financing-disguised-as-operating cash, big-bath restructurings) are tested and cleared, not silently dropped.

The verdict, and the one thing that would change it

Top two concerns. (1) Earnings quality: the last two years of reported growth were carried by a non-recurring disposal gain and a swing from reserve build to reserve release, not by underlying revenue, which was essentially flat ex-gain [11]. (2) Reserve adequacy: allowance coverage fell to 10.06% — the lowest since CECL — in the same year the loss-forecasting methodology was changed mid-cycle [3][7].

Cleanest offsetting evidence. Cash-flow presentation is textbook-clean — the $1,069M Pets Best gain is fully reversed out of operating cash, loan growth sits in investing, and securitization funding sits in financing [4]. ICFR is effective with no material weakness, and the actual credit data improved: 30+ delinquency fell to 4.49% and the net charge-off rate fell 66bps to 5.65%, which genuinely supports some reserve relief [1][10].

The one data point that would move the grade. A re-acceleration of net charge-offs or delinquencies into 2026 while coverage stays near 10% would convert the reserve-release yellow flag into a red one — it would mean earnings were borrowed from the future. Sustained credit improvement with coverage holding would push this toward Clean.

Where the earnings growth actually came from

The starting point of any Synchrony forensic read: net profit has not grown. Net earnings have round-tripped between $1.4B (2020 COVID build) and $4.2B (2021 COVID release) and sit at $3.55B in 2025 — below 2019 and 2021. Reported per-share growth is a buyback artifact: the diluted share count has fallen 55% since 2016, and roughly $19.7B of stock has been repurchased over that span.

Source: derived from reported financials, FY2019–FY2025; net earnings tie to the FY2025 10-K Statement of Earnings [1].

The table below separates the two stories — a flat profit pool, and a shrinking share count that manufactures EPS growth. This is not a shenanigan in the accounting sense; it is disclosed and legal. But a PM underwriting "double-digit EPS growth" should know the engine is the buyback, not the business.

Source: derived from reported financials, FY2019–FY2025 (net earnings, EPS and share counts as reported; buybacks per Consolidated Statements of Cash Flows) [4].

The FY2024 one-time gain (EM3 / EM7). Reported FY2024 net earnings rose 56.3% to $3.5B — but management itself attributes that almost entirely to the $802M after-tax gain on the sale of Pets Best [2]. The $1.07B pre-tax gain landed in the "Other" line of Other Income, which ballooned to $1,315M in 2024 and collapsed back to $295M in 2025 once the gain rolled off [6]. To Synchrony's credit, it removes this gain in its own "adjusted" figures rather than burying it — its non-GAAP framing is conservative here, lowering rather than flattering reported results.

The FY2025 reserve release (EM3 / EM6). With the gain gone, FY2025's modest +1.5% growth came from the provision line: provision for credit losses fell $1.5B to $5,225M, "primarily driven by lower net charge-offs, as well as a reserve release in the current year as compared to a reserve build in the prior year" [1]. On the Q1 call, the CFO quantified the swing precisely: a $97M reserve release versus a prior-year $299M build (which itself included a $190M day-one reserve on the acquired Ally Lending book) — and conceded net revenue fell 23%, "essentially flat" excluding the Pets Best gain [11].

Reserve timing manufactures the earnings cycle

This is the single most important forensic lens for Synchrony. Because provision is the largest swing item in the P&L, the direction of the reserve — build or release — is the main thing moving reported earnings. The chart below derives the annual reserve build/(release) from the change in the allowance balance. The pattern is unmistakable: a $1.6B release in 2021 (the COVID over-reserve unwinding) inflated that year; builds in 2022–2024 depressed earnings; and a $487M release in 2025 added it back.

Source: derived from the change in Allowance for Credit Losses, FY2020–FY2025 balances; allowance levels per FY2025 and FY2022 10-K statistical data [3][13].

Synchrony does not hide this — in FY2022 it told investors net earnings fell 28.5% "primarily driven by increases in provision for credit losses, primarily due to reserve reductions in the prior year" [12]. And its compensation committee neutralizes reserve releases in incentive metrics, which is itself an admission that the company treats releases as non-core. The forensic point is not deception; it is that a reader who annualizes any single year's earnings is mis-reading the through-cycle profit power, because reserve timing is doing the work.

Coverage fell to a CECL low — alongside a model change

The reserve-release tailwind is only a yellow flag (not red) because the underlying credit data improved. But the magnitude and the mechanism deserve scrutiny. Allowance coverage — the allowance as a percent of period-end receivables — has declined to 10.06%, the lowest reading since CECL adoption, even though it remains well above the 5.65% annual net charge-off rate.

Source: FY2025 10-K (2023–2025) and FY2022 10-K (2020–2022), Other Financial and Statistical Data [3][13].

The complication is timing. During Q1 2025, Synchrony changed its ACL methodology — replacing an "enhanced migration analysis" with a "statistical, account-level model" of probability-of-default and exposure-at-default, and switching its mean-reversion approach from a weighted method to straight-line [7][8]. A model change applied prospectively in the same year coverage hits a multi-year low is exactly the kind of judgment that warrants underwriting — and the auditor agrees, naming the ACL the sole critical audit matter and citing "significant measurement uncertainty" [9]. The honest read: the release is directionally defensible (delinquency and NCO both fell, from prior credit actions [14]), but the company has now both released reserves and re-modelled the estimate in the same period, so the quality of the coverage figure is harder to verify than in prior years.

Cash-flow quality: clean, and here is the mechanism

The brief's discipline — never accept strong CFO at face value — matters here because Synchrony's operating cash flow ($9.85B in FY2025) is nearly 3x net income, which on a naïve screen looks like aggressive cash generation. It is not. The mechanism is the provision add-back: provision for credit losses is a non-cash charge, so the full $5,225M is added back to net earnings in operating cash flow. That single line explains the bulk of the CFO-over-NI gap, and it is structural to lending, not a working-capital lifeline.

Source: FY2025 10-K, Consolidated Statements of Cash Flows [4].

Three classification tests come back clean. (CF1) Securitized-debt issuance ($2,242M) and repayment sit in financing, not operating. (CF2) The net increase in loan receivables (-$4,606M) sits in investing, where lending outflows belong. (CF3/CF4) The $1,069M Pets Best gain is fully reversed out of operating cash, and the $491M of sale proceeds is recorded in investing [4]. There is no evidence of financing inflows dressed as operating cash, and no acquisition/disposal distortion of CFO. Free cash flow and acquisition-adjusted FCF are not meaningful constructs for a deposit-funded lender (loan growth is the "investment," funded by deposits), so the relevant cash test is the classification integrity above — and it passes.

Soft assets: capitalized software is outrunning amortization

The one earnings-quality flag on the expense side is capitalized software (EM4 / CF2). Synchrony capitalizes internal software development and amortizes it over roughly five years. In FY2025 it added $765M of intangibles (mostly capitalized software) against just $363M of intangible amortization — additions are running at ~2.1x the expense. Gross capitalized software has reached $3,072M and the net balance jumped 45% in one year, to $1,151M.

Sources: FY2025 10-K Note 7 (FY2025: additions $765M, amortization $363M, net $1,151M) [5]; FY2024 10-K Note 7 (FY2024 additions $363M) [16]; FY2023 10-K Note 7 (FY2023 additions $392M) [17].

When additions outrun amortization, the net asset grows and current technology spend is deferred off the income statement. The $359M one-year increase in the net balance is the rough size of the expense deferral — about 8% of pre-tax income. It is disclosed, it is a legitimate GAAP policy, and the amortization tail (rising to a projected $378M in 2026) will eventually catch up [5]. But the direction — accelerating capitalization into flat revenue — is a yellow flag worth tracking, because it flatters reported "other expense" and the efficiency ratio today at the cost of higher amortization later.

Metric hygiene and revenue recognition

Non-GAAP (KM1). Synchrony's adjusted measures are unusually well-behaved: the marquee adjustment (removing the Pets Best gain) reduces reported earnings, and the efficiency ratio it features in incentive comp is consistently defined [3]. The hygiene concern is not aggressive add-backs; it is the recurring "transaction-related activity and other notable items" framing that risks normalizing serial one-offs, and the per-share optics already covered — adjusted/headline EPS growth that rests on the buyback, not profit.

Revenue recognition (EM1 / EM2). As a card lender, Synchrony recognizes interest income and continues to accrue finance charges on delinquent accounts up to 180 days past due, then reverses uncollected accrued interest at charge-off — a standard, disclosed policy with no sign of pull-forward beyond the norm. Revenue has no related-party content post-GE separation; it is interest and fees from cardholders plus interchange, with retailer share arrangements (a contra-revenue profit-share with partners) and loyalty costs netted transparently [6]. Interest and fee income was essentially flat in FY2025 (-0.2%) against a -0.9% move in receivables — receivables are not racing ahead of revenue [1]. Both EM1 and EM2 come back with no clear evidence of manipulation.

Breeding ground: governance dampens the accounting flags

The structural conditions are, on balance, dampening rather than amplifying. Synchrony is no longer a controlled company — its GE parentage ended a decade ago — so the original carve-out related-party risk has run off. Internal control over financial reporting is effective with no material weakness, and the financial statements carry an unqualified opinion from KPMG [10]. KPMG has served since the IPO — long tenure is the one auditor-side yellow flag — and stockholders are again asked to ratify it for 2026 [15]. The genuine breeding-ground tension is incentive design: compensation tied to EPS, ROTCE and an efficiency ratio in a company aggressively retiring shares creates a motive to favor per-share optics — which is consistent with (though not proof of) the buyback-driven EPS pattern. There is no founder/promoter dominance, no auditor resignation, no late filing, and no restatement.

The 13-category shenanigans scorecard

Source: synthesized from the evidence cited throughout, anchored to FY2025 10-K Summary Highlights, Statistical Data, Cash Flows, Note 7, Critical Accounting Estimates and Critical Audit Matter [1][3][4][5][7][9].

What to underwrite next

The accounting risk here is a valuation/quality-of-earnings haircut, not a thesis breaker. Underwrite the reported numbers, but discount the growth narrative and watch five things into the next two filings:

1. Allowance coverage vs. credit direction. Track the coverage ratio (10.06% at FY25) against 30+ delinquency (4.49%) and the NCO rate (5.65%). The downgrade signal: coverage flat-to-down while delinquency/NCO turn back up — that means the FY25 release borrowed from the future [3]. The upgrade signal: coverage holds while credit keeps improving.

2. The ACL model change's second year. FY2026 is the first full year on the new account-level PD/EAD model. Compare the disclosed qualitative overlay and macro-scenario weighting to FY2025; a quiet further reduction in the qualitative buffer would be a flag [7].

3. Capitalized software net balance. Watch Note 7: if additions keep running near 2x amortization and the net balance keeps compounding, the deferred-expense tail grows. A reversal (additions falling toward amortization) would be a positive [5].

4. Core revenue ex-reserve, ex-gain. Net interest income after RSA and provision is the real engine. Strip reserve releases and disposal gains each quarter and judge whether that line grows. FY2025 was essentially flat ex-gain [11].

5. Pace of buyback vs. profit. If net earnings stay flat while EPS keeps rising on shrinking share count, size the position to the profit pool, not the per-share line [4].

Bottom line. Synchrony's financial statements are credible — audited clean, cash flow honestly presented, disclosures thorough. The forensic issue is not whether the numbers are real, but whether the reported growth is. It is not: the last two years leaned on a one-time gain and a reserve release, the underlying profit pool is flat, and EPS growth is a buyback. None of that is a shenanigan in the fraud sense — it is a quality-of-earnings discount. Treat reported EPS growth as suspect, hold the allowance to account next year, and the accounting risk stays a footnote-to-haircut rather than a reason to avoid the name.