Financial Shenanigans

Financial Shenanigans — Molina Healthcare, Inc. (MOH)

Molina is not being accused of fraud here. The accounting that matters at Molina is judgmental — incurred-but-not-paid medical claim reserves (IBNP), risk-corridor and minimum-MLR accruals, Marketplace risk-adjustment payables, prior-year reserve releases, and what gets called "non-recurring" — and the past four quarters have stressed every one of those judgments at once. The forensic question is whether the reported numbers still faithfully describe the underlying economics, or whether multi-year tailwinds were being booked through reserves, "amounts due government agencies," and non-GAAP add-backs in a way that is now visibly unwinding.

Forensic Risk Score (0-100)

Risk Band

Elevated

Red Flags

Yellow Flags

3y CFO / Net Income

0.65

3y FCF / Net Income

0.54

FY2025 Accrual Ratio

6.5%

FY2025 Non-GAAP Gap (Adj NI / GAAP NI − 1)

23.6%

FY2025 Operating Cash Flow ($)

-$535,000,000

FY2025 Free Cash Flow ($)

-$636,000,000

FY2025 Prior-Year Reserve Release ($)

$98,000,000

Verdict

Forensic Risk Score: 62 / 100 — Elevated. Reported FY2025 GAAP earnings of $472 million are not obviously misstated, but the combination of (1) operating cash flow swinging from +$1,662 million in FY2023 to negative $535 million in FY2025 [1], (2) a $98 million total prior-year reserve release in FY2025 versus $675 million in FY2024 and $427 million in FY2023 [2][3], (3) a $93 million intangibles impairment booked in Q1 2026 immediately after the FY2025 10-K added a new goodwill/intangibles impairment risk factor [4][5], and (4) a covenant amendment cutting required interest coverage from 3.00x to 1.75x in February 2026 [6] means the prior multi-year earnings stream rested on reserves, risk-corridor payables and adjusted-EPS framing that are now visibly compressing.

Top two concerns: (a) CFO quality (CF4): management itself has now stated on the FY2026 first-quarter call that "operating cash flow swings a lot as we do accruals for risk adjustment for corridors, we hold those accruals. Maybe we don't pay them down for a year or two" [7] — i.e. the prior $2.1 billion / $1.9 billion CFO years were structural payable build-ups, not recurring cash generation; and (b) reserve and non-GAAP hygiene (EM6, KM1): adjusted net income of $584 million on $472 million of GAAP net income in FY2025 was reached after adding back $91 million of intangible amortization, $55 million of acquisition-related expense and other items [8], with "acquisition-related expenses" recurring every year since 2021 [9] — and the same Adjusted Net Income per Diluted Share is the only metric the company names as most important in linking executive pay to performance [10].

The cleanest piece of offsetting evidence: the auditor's Critical Audit Matter remained scoped only to the IBNP claims reserve in every annual report from FY2021–FY2025, internal control over financial reporting was assessed effective with no material weakness, and the compensation committee withheld the 2025 NEO short-term cash bonus when adjusted EPS fell below threshold [11] — none of this looks like a control-failure or pay-grab environment.

One data point that would change the grade: if Q2 and Q3 FY2026 operating cash flow comes in modestly positive on a normalized basis (i.e. with risk-corridor payables and Marketplace risk-adjustment payables roughly flat) and Days in Claims Payable returns to the 48–50 range, the Elevated grade rolls down to Watch. If, conversely, the FY2026 risk-adjustment payable continues to build at the Q1 pace ($630 million payable vs. $112 million receivable, a net $518 million payable [12]) and DCP stays below 46, the grade tips to High.

The 13-category shenanigans scorecard

This is the standardized accounting-quality map. Read each row as "is there evidence Molina did this?" The evidence column is the specific line item, period and disclosure; the prose around the table carries the page citations.

No Results

Three categories are red: EM5 under-reserving, EM6 smoothing, CF4 unsustainable CFO, and KM1 non-GAAP hygiene. Six are yellow; three pass cleanly with no clear evidence. The case is not that any single ratio is shocking — it is that the four red rows trace to the same underlying mechanism: corridor and reserve accruals that built up in 2020–2023 and have unwound through 2024–2025, leaving recent reported earnings without the cushion they had before.

The cash-flow mechanism: $2.1B → $-535M is a payable unwind, not an earnings collapse

The single most important forensic chart on this name is the multi-year walk of operating cash flow against net income. CFO ran at 3.22x net income in FY2021 and 2.82x in FY2020 — both of which precede an enormous build in "Amounts due government agencies" as COVID-era state-level risk corridors and minimum-MLR rebates were accrued but not yet paid. CFO then converges with net income in FY2022, sharply exceeds it again in FY2023 (1.52x), and finally goes negative in FY2025 (–1.13x) as those same payables clear cash. Management said it explicitly on the Q1 FY2026 call: "Operating cash flow swings a lot as we do accruals for risk adjustment for corridors, we hold those accruals. Maybe we don't pay them down for a year or two" [7].

The FY2023 10-K MD&A and the FY2025 10-K MD&A explain the swing as "timing differences in settlement of government agency receivables and payables, including settlements for Medicaid minimum MLR and medical cost corridors and Marketplace risk adjustment payables" [1]. The Q3 FY2025 10-Q shows the nine-month FY2025 CFO outflow of $237 million driven by a $467 million release of "Amounts due government agencies" and a $157 million release of medical claims and benefits payable [13]. Those payables had been built in prior years — Marketplace risk-adjustment net payable was $411 million at year-end 2025 versus $98 million at year-end 2024 [14], and rose to a $518 million net payable by March 2026 [12]. The Q1 FY2026 +$1,082 million CFO spike is itself driven by exactly the same lever in the other direction (the rebuilding of risk-adjustment payables alongside deferred revenue that jumped from $66 million to $401 million inside one quarter) [15].

This is the mechanism behind CF4. The 5-year cumulative CFO/NI of 1.11x flatters Molina: it averages a $4.7 billion working-capital lifeline in FY2020–FY2023 against a $1.7 billion working-capital drain in FY2024–FY2025. The recurring cash-generation underneath is closer to net income, and possibly somewhat below it once acquisition spend is netted.

Prior-year development: the reserve cushion is empty

The reserve roll-forward in Note 10 ("Medical Claims and Benefits Payable") is the single most informative disclosure in any managed-care 10-K. It shows how much of current-year reported medical cost was actually prior-year liabilities being released. For FY2023 the consolidated prior-year favorable development was $(427) million [16]; for FY2024 it jumped to $(675) million ($(611) million from Medicaid, $(61) million from Medicare, $(3) million from Marketplace) [3]; for FY2025 it collapsed to $(98) million in aggregate, with Marketplace actually showing an unfavorable $61 million [2]. Against an FY2024 GAAP net income of $1,179 million, the FY2024 PYD release alone was the equivalent of 57% of net income at a 23.7% federal-equivalent tax rate.

Management's own FY2025 disclosure is forensically explicit: "The impact of prior year reserve development in 2025 was partially absorbed by minimum MLRs and medical cost corridors and was ultimately not material to our consolidated MCR" [17]. That sentence — repeated almost verbatim in Q1 FY2026 [4] — is the smoothing-mechanism admission. Reserves released in good years that would otherwise have flowed to income were absorbed by paying corridors back; reserves over-released in bad years are similarly capped because the corridor / minimum-MLR rebate doesn't fully credit them either. This is not necessarily abuse — it is the structure of state Medicaid contracts. But it does mean the historical PYD pattern is not a measure of conservatism; it is a measure of how much cushion was available to absorb adverse trend. In FY2024 there was $675 million of cushion. In FY2025 there was $98 million. In Q1 FY2026 management said the PYD impact was "mostly absorbed by minimum MLRs and medical cost corridors" [4], i.e. effectively zero.

The SEC has already walked Molina through this exact dynamic. In a December 2018 comment-letter response, Molina disclosed that 2017 unfavorable development was attributable to "inaccurate adjudication of provider claims at our Florida, Illinois, New Mexico and Puerto Rico health plans" and that the issues "existed at the time we made our December 31, 2016 liability estimate" [18]. In a February 2019 follow-up, the SEC explicitly asked Molina to explain why a restatement of the FY2016 reserves was not required under SAB 99, which the company defended on materiality grounds [19]. Days in Claims Payable was committed to be maintained in the 50–55 day range in the same correspondence cycle, but Q1 FY2026 DCP is 44 [20], below that historical floor for the fifth straight quarter.

Management consistently attributes the drop to "pass-through payments", "timing of payments at quarter end" and faster claims adjudication. That is plausible. It is also what someone would say if cushion were thinner. Either way, the historical 50–55 day commitment to the SEC has not held since FY2024.

The non-GAAP wedge: adjusted EPS is the company's pay metric

In FY2025 GAAP net income was $472 million ($8.92 per diluted share); adjusted net income was $584 million ($11.03 per diluted share), reached by adding back $91 million of intangible-asset amortization, $55 million of acquisition-related expenses and $1 million of "other", tax-affected at the company's marginal rate [8]. The FY2024 reconciliation showed $1,179 million GAAP net income against $1,308 million of adjusted, with similar add-backs of $83 million intangible amortization, $66 million acquisition-related and $16 million other [21]. The same line, "acquisition-related expenses", recurs in every published reconciliation: $93 million FY2021, $49 million FY2022, $7 million FY2023, $46 million in nine-month FY2024 [9]. When a line item appears every year, it is no longer "non-recurring".

The proxy is unambiguous about the incentive alignment: "Adjusted Net Income per Diluted Share" is the only measure the 2024 Pay vs. Performance disclosure names as most important in linking compensation actually paid to NEO performance [10]. The 2022 PSU grant vested at 149% of target on the basis of adjusted-EPS achievement of $46.31 over fiscal 2022 / 2023-2024 [22]. The 2025 LTI grant uses three-year cumulative adjusted EPS for FY2024-FY2025-FY2026 [23]. The forensic point is not that the company is dishonest about the gap (the reconciliations are clearly presented) — it is that the company has every incentive to (a) keep "acquisition-related" and "amortization" structurally above the line in the adjusted measure, and (b) hold the adjusted line up even when GAAP compresses. There is one mitigating data point: the compensation committee did not pay an FY2025 short-term cash bonus to the NEOs because adjusted EPS of $11.03 came in below threshold [11]. That is a meaningful check.

A separate KM1 detail: mid-2025 the headline margin metric in the earnings release moved off after-tax margin and onto pre-tax margin during a weaker year. Switching the denominator the year reported margins compress is the classic key-metric tell, even if each individual disclosure is internally consistent.

Big-bath behavior: real estate 2022, MAPD intangibles 2026

Two impairments inside three years line up with the EM7 pattern. In December 2022, Molina filed an Item 2.06 8-K announcing a non-cash, pre-tax impairment of approximately $200 million "attributable to leased space," explicitly noting "this charge will be recorded outside of adjusted net income" [24]. The FY2022 10-K MD&A confirms the recorded charge was $208 million [25]. In February 2026, the same management team announced the FY2025 results, signaled a MAPD product exit for 2027, and the Q1 FY2026 10-Q recorded a $93 million impairment charge related to our planned exit of the MAPD product in 2027 [4]. The FY2025 10-K had already added an explicit goodwill / finite-lived intangibles impairment risk factor immediately preceding the announcement [5], which is itself the kind of signposting the playbook flags.

Neither impairment is alarming in isolation — both are disclosed cleanly, both are tied to identifiable strategic decisions (permanent remote work; MAPD product exit), and both are arguably economically defensible. What raises the flag is the pattern: the two largest non-cash charges in five years were both routed to non-GAAP adjustments, both took place in or just after weak reported years, and the second one is now eating an entire acquisition's identifiable intangibles. The Bright HealthCare California Medicare deal closed in early 2024 with $295 million of cash and substantial intangibles; the MAPD impairment in Q1 FY2026 is the partial unwind of those intangibles.

Acquisition optics on cash flow (CF3)

Molina has been an acquirer every year since 2020: Magellan Complete Care 2020, Affinity New York Medicaid 2021, Cigna Texas / AgeWell 2022, My Choice Wisconsin 2023, Bright HealthCare California Medicare 2024, ConnectiCare 2025. The FY2025 reserve roll-forward shows $379 million of acquired medical-claims-and-benefits-payable balances on the year (mostly from ConnectiCare) [2]. Acquired claims liabilities are a real liability the buyer assumed, but the cash that came with them flows through operating activities as a working-capital pickup in the acquisition quarter — which means the first reported CFO post-deal is structurally flattered. Backing out cash used in business combinations (FY2025: $245 million, FY2024: $295 million plus $49 million of post-closing) gives a more honest read of recurring free cash generation [1]:

The chart shows what every acquisitive compounder hates: the post-acquisition FCF line is materially below the headline FCF line in three of the last five years, and is now meaningfully negative.

Earnings-quality tests that come back clean

Negative evidence builds trust. Of the 13 categories, three pass cleanly with no material concern:

EM1 (premature revenue): Premium revenue recognition is per-member-per-month, tied to entitlement to receive services, and is the model the entire managed-care industry uses. The FY2025 10-K policy note is unchanged from prior years and walks the reader through risk corridors, minimum MLRs, retroactive premium adjustments, and CMS Part D / Marketplace risk-adjustment separately [14]. There is no contract-asset / unbilled-receivable line of any size; receivables are 81% government receivables of $2,380 million [26] and the credit-loss allowance is described as "insignificant" because the counterparties are governmental.
EM2 (bogus revenue): Customer concentration is state Medicaid agencies and CMS. The Texas Medicaid contract alone is $5,735 million (18% of consolidated Medicaid premium) and the Washington contract is $4,194 million (13%) [27]. Related-party transactions: the 2026 proxy describes a formal review policy and does not disclose any related-person transactions of consequence after the Molina-family era; no parent-CEO / family-controlled-counterparty pattern remains [28].
CF1 (financing inflows recharacterized as operating): There is no disclosed factoring program, no receivables securitization, no supplier-finance program, and the credit agreement permits but does not record any such structure. The $850 million senior notes due 2031 issued in November 2025 are routed through financing activities cleanly [29].

Breeding ground — does the governance amplify or dampen the red flags?

This is where Molina earns some benefit of the doubt. The 2026 proxy shows an independent board (10 nominees, all independent except the CEO), an Audit Committee chaired by a CPA-credentialed director, and a formal related-person-transactions policy with annual D&O questionnaires [28]. Critical Audit Matter from PwC was scoped to IBNP claims reserves in FY2021, FY2022, FY2023, FY2024 and FY2025 — the auditor signals it knows the same area carries the most estimation risk, and the company has reported no material weakness in internal control over financial reporting. The compensation committee withheld the FY2025 NEO short-term cash bonus when adjusted EPS came in at $11.03 against threshold [11]. There is no founder / family voting dominance after the 2017 board turnover, and no "perfect" expectation-beating streak (the FY2025 adjusted EPS guidance was cut from $24.50 to $14 over a single year, with two downward revisions in print [30]).

What works the other way: (i) Adjusted Net Income per Diluted Share is the single named compensation metric in the Pay-vs-Performance table [10] — the company has every reason to push that line up; (ii) the 2025 proxy disclosed "non-recurring" performance-based one-time retention grants to Messrs. Zubretsky and Keim, which is the kind of "non-recurring" label that becomes recurring; (iii) Molina has been on the SEC's accounting-disclosure radar more than once on this exact subject area — IBNP reserves, prior-year development, real-estate impairment classification — between 2008 and 2019.

Net read: the breeding ground is healthier than typical for a "62" forensic score. The governance does not look like a fraud breeding ground; it does look like an environment where adjusted-EPS framing is structurally rewarded. That dampens the worst-case (intentional manipulation) but does not dampen EM5/EM6/KM1 — those red flags are about the chosen accounting policy and the chosen non-GAAP framing, not about the controls failing.

Leverage: covenant stress is the loudest yellow flag

At Q4 FY2025, debt was 3.7x trailing-12-month EBITDA and debt-to-cap was approximately 49% [29]. By Q1 FY2026 the ratio had spiked to 6.1x trailing-12-month EBITDA because of the EBITDA denominator collapse in Q1 [20]. On February 6, 2026 Molina amended its credit agreement to cut the required minimum interest-coverage ratio from 3.00:1.00 to 1.75:1.00 for the four quarters of fiscal 2026, then a stepped recovery to 2.75:1.00 by Q3 2027 [6]. A covenant amendment to head off a breach is a meaningful disclosure — it is the lender writing down in black ink that current EBITDA-to-interest is approaching the contractual line.

What to underwrite next

This is a forensic page for an investor underwriting position size, not for a litigator. The five items to watch next:

FY2026 Q2 and Q3 operating cash flow — does the $1.1 billion Q1 build hold, or is it pure timing? Watch the Marketplace risk-adjustment payable line (now $518 million net payable [12]) and the Amounts-due-government-agencies line.
PYD in FY2026 — if reported PYD turns unfavorable in any segment, prior under-reserve thesis is confirmed. The pattern in FY2025 was already unfavorable in Marketplace ($61 million) [2].
DCP trajectory — return toward 48–50 settles the under-reserve question. Drift further below 44 is a downgrade trigger.
Goodwill / intangibles impairment cadence — the FY2025 10-K added the goodwill impairment risk factor [5]; ConnectiCare goodwill ($287 million added in FY2025 [31]) is the next candidate if the Marketplace book deteriorates.
Interest-coverage covenant — a second amendment, or compliance with 1.75x being meaningfully tight on the Q1/Q2 2026 print, is a thesis-breaking signal. The Q1 2026 interest expense was running at roughly $30 million quarterly against EBITDA that has compressed sharply.

Signal that downgrades the grade (toward Watch): Q2 FY2026 CFO comes in positive on a normalized basis; DCP returns to 47+; reserve roll-forward shows neutral-to-favorable PYD in Medicaid; no further impairment charges.

Signal that upgrades the grade (toward High): another covenant amendment; PYD unfavorable in two segments; cumulative Marketplace risk-adjustment net payable exceeds $750 million; or auditor scope adds a second Critical Audit Matter.

Should this work affect position sizing? Yes. The forensic risk is not "earnings are fake" — the GAAP numbers tie cleanly and the auditor signs them. The risk is that the prior-year trajectory of adjusted EPS was supported by reserve cushion and corridor accruals that have now been spent, and that the headline non-GAAP framing makes the deterioration look less severe than it is at the cash and reserve level. For a long position, this argues for a position-sizing limiter (smaller than the headline P/E would suggest) and a valuation haircut on the embedded-earnings narrative ($8.65 of "embedded earnings" framed on the Q3 FY2025 call [30] should be discounted, not added). It does not, on this evidence, support calling the company unethical or the audit opinion unreliable.

References

Generated Jun 19, 2026 10:03 UTC