Damage Math

Damage Math

Cigna's price fell about a third from its September-2024 peak, yet consensus adjusted EPS never fell — it grew, $27.33 in FY2024 [1] to $29.84 in FY2025 [2] toward $30.41 for FY2026E. The trigger was a stop-loss mispricing worth roughly 100 bps of one segment's margin. Under conservative assumptions the problem destroys $0.7–9B of net present value against $27–38B of erased market value. The independent trial puts the probability the impairment is temporary at 0.62.

The near-term hit: consensus earnings barely moved

The pattern Ruchir's framework hunts is a company where near-term earnings were cut and the stock was anchored to the cut. Centene is the archetype — a roughly two-thirds guidance cut and a roughly two-thirds price fall. Cigna is a weaker version of that setup on the earnings axis, because the near-term earnings line did not break.

Adjusted EPS grew straight through the drawdown. FY2023 landed at $25.09 [3], FY2024 at $27.33 — up 9% but, in management's own words, "short of our outlook" [4] — and FY2025 at $29.84, up another 9.2% [5]. The one visible dent was a single quarter: Q4 2024 normalized EPS of $6.64 against a $7.82 consensus, a −15.1% miss of about $1.18 per share, the only normalized-EPS miss in the last sixteen quarters (per S&P Capital IQ consensus). The full-year number kept rising.

Actuals: Q4 2023–Q4 2025 earnings calls [6][7][8]. FY2026E–FY2029E: consensus, S&P Capital IQ (data/sp/estimates.json).

What did fall was the growth rate, for one year. Management guided FY2025 to "at least $29.50 per share" at the Q4 2024 call [9] — a step down from the company's stated 10–13% long-term algorithm — then set FY2026 at "at least $30.25 per share" [10], later raised to "at least $30.35" after Q1 2026 [11]. At $30.41, FY2026 consensus is only about 1.9% above FY2025 — the single year of below-algorithm growth. FY2027 consensus of $33.48 resumes roughly 10% growth. On the vendor record, the out-year lines have barely moved: over the trailing six months FY2027 normalized EPS went from $33.51 to $33.48 and FY2028 from $37.00 to $37.01 — flat, within a percent (S&P Capital IQ point-in-time consensus). The consensus tape shows a one-year pause, not a permanent reset.

That is the numerator of this tab, and it is small: a one-quarter miss of about $1.18 per share, plus one year of growth running roughly 8–9 points below the algorithm before it re-accelerates on consensus. The Dislocation tab covers what the price did over the same window.

The price and enterprise value: a third of the equity, gone and partly back

Against that near-flat earnings line, the equity moved a great deal. The stock peaked at $366.85 on 16 September 2024, troughed at $244.41 on 31 October 2025 — a 33.4% decline over 410 days — and sits at $284.85 as of 22 July 2026 (derived: fit_features.capitulation_gauge). Translated to market capitalization on the share counts of the day, that is roughly $104B at the peak, about $66B at the trough, and $76.5B now (derived: fit_features.market_cap; share counts from fit_features.share_count_trend).

Prices: fit_features.capitulation_gauge (data/prices/daily.json); market cap = price × shares outstanding of the period (fit_features.share_count_trend, market_cap).

The equity lost about $38B from peak to trough (−36%, slightly more than the −33.4% price move because the share count also fell as buybacks continued), and it sits roughly $27.5B below the peak today (−26%). Enterprise value moved almost entirely with the equity: net debt is a stable-to-declining ~$24B (consensus net debt $23.9B for FY2025, guided down toward $18.7B for FY2026E, per S&P Capital IQ), so EV fell by essentially the same $27–38B as the market cap, from roughly $130B toward $100B. Debt is not the story here; the equity is.

Put the two sides together and the mismatch is the whole point of this tab: consensus FY2026 adjusted EPS rose about 2% while the market cap fell about 26% from the peak. Almost none of the price move is an earnings move.

The move was a multiple, not an earnings collapse

Holding forward earnings fixed at today's FY2026 consensus of $30.41 isolates what actually changed — the multiple the market was willing to pay.

Derived: peak/trough/current price ÷ FY2026 consensus adjusted EPS of $30.41 (data/sp/estimates.json; prices from fit_features.capitulation_gauge). Forward EPS at the peak was itself near $30, so the ratio characterizes a multiple move.

The price traveled from about 12.1x at the peak to 8.0x at the trough to 9.4x now, on a forward EPS number that has been near $30 throughout. A useful cross-check runs the multiple backward into an implied terminal-growth rate: at the current $76.5B market cap, on ~$8.0B of normalized after-tax earnings (FY2025 net income normalized, data/sp/estimates.json) and a 9% discount rate, the price implies adjusted earnings shrinking about 1.3% a year in perpetuity. At the peak market cap of ~$104B, the same arithmetic implied earnings growing about 1.2% a year. The de-rating swung the market's implied perpetual-growth assumption by roughly 2.5 points — from mild growth to mild permanent decline — on a business whose revenue grew 11% in FY2025 and whose adjusted EPS grew 9%.

The NPV arithmetic: how much value could the problem plausibly destroy

The question is whether a hit this size can justify $27–38B of lost value. A transparent two-scenario DCF-lite answers it. The assumptions are deliberately simple and stated so a reader can reproduce every number.

Cigna Healthcare booked $47.16B of adjusted revenue and $4.15B of adjusted operating income in FY2025 [12]; 100 bps of margin on that revenue base is about $0.47B pre-tax. Evernorth's Q1 2026 Pharmacy Benefit Services pre-tax income fell from $544M to $394M — about −28% — which annualizes to roughly $0.6B of pre-tax income [13]. Those are the two hit sizes; the rest is discounting.

Derived: NPV = after-tax hit discounted at 9%. Temporary = two years discounted then recovery; permanent = level-shift perpetuity (after-tax hit ÷ 0.09). Hit sizes from FY2025 10-K segment revenue [14] and Q1 2026 10-Q [15].

The temporary reading destroys about $0.65B of value — under 1% of the market cap. The permanent reading, treating the PBM compression as a level-shift that never reverses, destroys about $5.3B; pushed to a severe $1.0B-a-year permanent hit, about $8.8B, or roughly 12% of the market cap. Set those against the price damage and the gap is the finding of the tab.

Derived: price damage from market-cap change (fit_features.market_cap, capitulation_gauge); NPV damage from the two-scenario DCF-lite above.

The gap is real and large. It is not, on its own, a verdict — a gap only becomes an opportunity if the impairment is in fact temporary, and that question is not settled by this arithmetic. It is settled, as fairly as the profile can, by an adversarial trial.

The trial: temporary versus permanent, both cases at full strength

Two opposing briefs argued the diagnosis from the same corpus; three blind judges ruled. Both cases have genuine evidence, and neither is a straw man.

Three independent judges, reading the briefs in randomized order, put the probability the impairment is temporary at 0.62 (mean of the panel 0.63; individual judges 0.59, 0.62, and 0.69; spread 0.10). Order stability was tight — the two reading orders differed by 0.02 — and the ruling was not flagged as contested (ruchir/trial/tally.json). This is the diagnosis probability the report carries; the arithmetic above cannot override it, and does not try to. The reading leans temporary, but with roughly a 38% weight on the permanent case that the PBM margin trend is genuinely re-basing lower — a weight the Durability tab tests on the 8–20 year horizon.

The panel also fixed what would flip the ruling toward permanent: FY2026 Cigna Healthcare MCR printing above 84.7% or adjusted EPS below $30.25 (repricing failed to stick); Evernorth or Pharmacy Benefit Services pre-tax income declining year-over-year beyond disclosed transition costs; PBM profit per claim staying below the 2025 base after rebate remittance and Part D delinking take full effect post-2028; or medical customers falling materially below ~18M beyond the deliberate exchange exit.

Which line broke, and whether it self-corrects

Two distinct drivers sit behind the hit, and they carry different recovery mechanisms.

Stop-loss (Cigna Healthcare) — mechanical repricing. Stop-loss is a one-year product priced annually; the 2024 cost spike was locked into 2025 pricing before it was visible, then addressed at the next renewal. Management's ~100 bps recapture is scheduled across 2026–2027 on that renewal cadence [30], and the FY2026 MCR guide of 83.7–84.7% already incorporates the pricing actions [31]. This is the piece that self-corrects, and it is the majority of the near-term earnings dent.

Evernorth / Pharmacy Benefit Services — the contested driver. Here the recovery mechanism is genuinely in question. Management frames the Signature transition as margin-neutral with 97%+ retention [32], but the segment's margin has fallen from 5.3% to 3.1% since 2019 [33], Q1 2026 PBS income fell 28% on higher revenue [34], and the 2028 rebate-remittance and Part D delinking rules are structural, not cyclical [35]. If this driver re-bases permanently, it is the permanent-case scenario in the DCF-lite above — worth roughly $5–9B of NPV, still well short of the price damage, but a genuine and structural reduction rather than a calendar artifact. The Durability tab carries the long-horizon read on this line, and the framework's adjusted-FCF yield — which strips stock compensation and average acquisition spend from the cash the buyback flywheel runs on — is built in Yield.

The bottom line of the arithmetic: the price destroyed on the order of $27–38B of market value; the problem, read conservatively and even at its permanent worst, plausibly destroys $0.7–9B of net present value. A gap of roughly $19–27B separates the two. Whether that gap is an opportunity turns on the diagnosis — and the trial puts the probability it is temporary at 0.62, not a certainty.