Clock

What has to happen, and when

The re-rating mechanism at Oscar Health is the ACA individual-market repricing cycle, and the record shows it has already fired: a roughly 28% weighted average rate increase for plan year 2026 filed across states covering close to 99% of membership [1], a first-quarter 2026 print of $2.07 per diluted share against $1.10 consensus [2], and a share price 161% above its 30 March 2026 trough. Long-dated listed options run to January 2028 at implied volatility near 78%.

The repricing mechanism

Oscar's economics reset on an annual clock by construction. The company elects to participate in each individual market on an annual basis, and its premium rates and specific rate changes require approval from state and federal regulators under the ACA [3]. That is the mechanism the framework's healthcare-forecasting-error pattern looks for: a cost miss in one plan year is repriced into the next year's premiums, industry-wide, on a calendar the regulator enforces.

The 2025 miss and the 2026 repricing are both on the record. Market morbidity stepped up across the industry as Medicaid redetermination lives entered the exchanges; Oscar's FY2025 loss from operations was $396 million [4]. Management refiled 2026 rates in states covering close to 99% of membership at a roughly 28% weighted average increase, explicitly reflecting elevated trend, higher 2025 market morbidity, the expiration of the enhanced premium tax credits, and CMS program-integrity initiatives [5]. Guidance introduced on 10 February 2026 put FY2026 revenue at $18.7 billion to $19 billion, up 61% year over year at the midpoint [6], and a swing of nearly $750 million in earnings from operations at the midpoint [7].

The second mechanism was a feared event whose consequence did not arrive at the size the market priced. The enhanced premium tax credits did expire at the end of 2025 [8], and Oscar had priced 2026 assuming exactly that [9]. What followed was membership of 3.2 million at 31 March 2026, up 56% year over year, and approximately 3 million paid members at 1 April with payment rates consistent year over year and modestly favorable to plan despite the subsidy sunset [10]. Against a company estimate of 20% to 30% market contraction [11], the first Wakely read tracked in line to favorable [12].

Both mechanisms are visible in one quarter's arithmetic. First-quarter 2026 revenue rose 53% to $4.6 billion, the medical loss ratio improved 490 basis points to 70.5% with utilization largely in line with expectations, and earnings from operations reached $704 million [13]. Net income was approximately $679 million, or $2.07 per diluted share, the highest in the company's history, and full-year guidance was reaffirmed [14].

The third mechanism the framework looks for — a shrinking share count — is absent here. Oscar's share count has risen every year on the record, with a five-year compound growth rate of 55.1%, and the feature file records the trend as rising; the detail sits in Self-Help.

What is still on the calendar

Sources: Q2 2026 call date, company announcement of 13 July 2026 [15]; Investor Day date and the 2027 pricing-cycle exchange, Q1 FY2026 transcript [16] [17]; Q2 as the first claims-based morbidity read [18] [19]; OEP window, litigation and OBBBA, FY2025 Form 10-K [20]; annual rate approval [21]; CMS June settlement cycle [22].

The nearest of these is nine days out. Management framed the second quarter as the first quarter in which claims data, rather than demographic proxies, shows what 2026 market morbidity actually is [23], and named the Wakely report and risk adjustment as the variables that could still move the 2026 outlook [24]. The risk-adjustment accrual was booked at 24.5% of premium in the first quarter [25] against a full-year expectation of approximately 20% [26], with none of the observed market morbidity favorability recognised [27] — so the mechanism has an identified, dated release valve rather than a hoped-for one.

Base rates from OSCR's own history

The price record runs from the 3 March 2021 listing to 27 July 2026 — 1,357 trading days. Segmenting it by 35% swing reversals produces thirteen distinct drawdown episodes, which is roughly one every five months.

Source: derived from the run's daily closing-price series, 3 March 2021 to 27 July 2026; episodes segmented at a 35% swing-reversal threshold on closing prices, company filings as reported.

The arithmetic a skeptic can recompute: median depth −42.1%, median 70 calendar days from peak to trough, and — across the twelve episodes that regained their prior peak — a median 170 days from trough back to that peak and 256 days for the full round trip. The tail is wide. Four episodes ran 50% or deeper, all of them in the 2021–22 de-rating, and their trough-to-prior-peak recoveries were 166, 574, 830 and 1,772 days, a median of 702 days. The shortest recovery in the whole set was five days.

Source: derived from the run's daily closing-price series, 3 March 2021 to 27 July 2026; month-end close against the running maximum close since listing, company filings as reported.

Two facts sit inside that shape. The stock has spent its entire listed life below its 10 March 2021 close of $36.77, and at $28.34 on 27 July 2026 it is still 22.9% below it — so at the whole-history level a gap remains. But measured against the episode the fit-feature file defines as the current drawdown — the 19 September 2024 peak of $23.27 to the 30 March 2026 trough of $10.85, a depth of 53.4% over 557 days, on a volume spike of 15.5x the pre-peak median — that gap is closed. The $23.27 peak was regained on 12 May 2026, 43 days after the trough and six days after the Q1 print, and the price is now 21.8% above it. A swing-based segmentation of the same series splits the feature file's single 557-day episode into three separate legs, the last of which fell 40.3% and retraced in 31 days; the two readings differ in bookkeeping, not in direction.

The current episode therefore has no precedent in this name's history in one specific respect: no prior drawdown of 50% or more retraced in anything close to 43 days. The four that did so took between 166 and 1,772 days.

The 18-month read

Re-recognition within 18 to 24 months is not the open question here, because the record shows it has already happened inside four months. From the 30 March 2026 trough at $10.85, the price reached $32.18 on 2 July 2026 — a 197% move in 94 days — and stands at $28.34, up 161% from the trough and 11.9% below that July high. The mechanism fired on schedule (2026 rates repriced ~28%, subsidy sunset absorbed, first quarter printed at a record), and the price moved with it. What remains is not a closing gap but a forward underwriting question about the 2027 cycle: whether pricing holds when the OEP window shortens to six weeks [28] and the stayed program-integrity provisions potentially return.

The strongest fact against that read is that the 2026 result is not yet proven. Full-year guidance implies earnings from operations of $250 million to $450 million against $704 million already booked in the first quarter, and consensus expects a fourth-quarter 2026 loss of $1.23 per share — the year's profit is a first-quarter phenomenon that the back half is expected to spend down. If the June Wakely claims data breaks against pricing, the mechanism unwinds inside the same year it fired. That is the falsifier this tab contributes: 2027 weighted average rate increases failing to cover realised 2026 morbidity, visible first in the 6 August 2026 and November 2026 prints and settled by the CMS final risk-score report in June 2027.

Street positioning and the printed quarter

Source: consensus estimates and target-price data as of 27 July 2026 (10 contributing targets; high $35, low $13), and the run's closing-price series.

The sell side has not capitulated in either direction, and it is behind the price. Eleven in-consensus recommendations split 3 outperform, 7 hold, 1 underperform, with no buy and no sell ratings — a 2.82 score that sits between outperform and hold. The mean target of $24.20 is 14.6% below the 27 July close and the median of $21.00 is 25.9% below it; the $35 high target is 23.5% above. A stock trading above the mean of the targets set on it is not a stock the sell side is waiting to discover.

What the sell side has done is revise. The estimate vintages tell that story more usefully than the ratings do.

Source: consensus normalized EPS estimate momentum, vintages dated 28 January, 28 April, 27 June and 27 July 2026.

FY2027 consensus normalized EPS has moved from $1.00 to $1.52 over six months, and FY2028 from $1.08 to $2.21 — a doubling. Consensus FY2027 revenue over the same window moved from $14.1 billion to $19.8 billion. The revisions are the re-recognition, and most of them landed between the January and April vintages, which brackets the March trough.

Source: reported quarterly EPS and consensus normalized EPS estimates; 2Q26 onward are consensus only, as of 27 July 2026.

Consensus itself puts the recovery in printed numbers in the quarter that has already printed. FY2026 normalized EPS consensus of $1.10 is the company's first profitable year, and 1Q26 alone delivered $2.07 of it. The remaining candidate quarters are 2Q26 on 6 August 2026 [29], where consensus wants $0.39 and the first claims-based morbidity read arrives, and 1Q27, where consensus wants $2.10 and where a second consecutive first-quarter record would establish the 2026 result as a cycle rather than a single reset year. On the 27 July close, consensus normalized EPS puts the shares at 25.8 times FY2026, 18.7 times FY2027 and 12.8 times FY2028 — the multiple already discounts two more years of the revision path holding.

Instrument facts

Listed options on OSCR exist across sixteen expiries, with the two longest dated 17 December 2027 and 21 January 2028 — 508 and 543 days beyond the 27 July 2026 quote, or 16.7 and 17.8 months. Both clear the framework's 12-month reference; the longest sits just short of 18 months.

Source: Cboe delayed option quotes for OSCR, timestamped 27 July 2026 21:49 UTC; expiries with under 1,000 contracts of open interest omitted from the table but included in the totals below. Implied volatility is open-interest-weighted across the listed strikes in each expiry.

Total open interest across the chain is 477,367 contracts, of which 130,498 — 27.3% — sits in the two expiries beyond twelve months. Concentration is high: the January 2027 and January 2028 expiries together hold 60.0% of all open interest. Trading in the long tenors on the quote date was thin: 47 contracts changed hands in the January 2028 expiry and 144 in December 2027, against 7,908 across the whole chain, roughly 88% of which was in expiries inside 60 days. Open interest in long-dated contracts exists; daily turnover in them does not.

Implied volatility is elevated against the framework's reference lines. Cboe published a 30-day implied volatility of 86.6% for OSCR on 27 July 2026. AlphaQuery's independently computed mean implied volatility for the same date reads 88.6% at 30 days and 77.4% at 180 days, against 30-day realised close-to-close volatility of 57.4%. Open-interest-weighted implied volatility in the December 2027 expiry is 77.8% and in January 2028 is 78.0%. The framework's own reference lines treat up to roughly 50–55 as acceptable and 60–70 as elevated; every tenor on this chain sits above 70, and the front end sits above 100.

Those are the instrument facts as of 27 July 2026, stated as facts. Long-dated contracts exist, so the framework's no-qualifying-LEAPS watchlist route does not apply; the implied-volatility level is what it is, and reads across to the damage arithmetic in Damage Math and the drawdown anatomy in Dislocation rather than to anything about expression.