CPA — Copa Holdings, S.A.

Copa Holdings, S.A. (NYSE: CPA) — deep-value analysis

Date: September 7, 2026 · Reference price: … (close 2026-09-06, yfinance) · Market cap: … · Shares: 40.82 mil (30.20 mil Class A outstanding + 10.94 mil Class B) · Reporting and trading currency: USD (identical — no conversion risk in the bridge)

Primary sources: 20-F FY2025 filed 26.02.2026 (SEC-Filings/CPA/CPA_20F_2026-02-26.txt), Q2 2026 earnings release (6-K from 07.08.2026, exhibit 99.1), data pack rapoarte/deep/data-pack-CPA-20260907.md, EIA Gulf Coast jet fuel spot price series, GBL tracker.

Executive summary

Copa Holdings operates the “Hub of the Americas” out of Panama City — 436 daily flights to 84 destinations in 32 countries, with a unique fleet of 131 Boeing 737s (125 at 12/31/2025, average age 10.2 years). It is, by construction, the most efficient airline in Latin America: operating margin of 22.6% in 2025 (… operating profit on … revenue, 20-F p. F-5), unit cost ex-fuel of 5.7 cents/ASM — a low-cost-carrier level achieved with a business-class product and 90.6% on-time performance. ROE 23.1%, CFO … in 2025, net debt … = 0.9× LTM EBITDA (the company’s own reconciliation, Q2 2026 release).

What broke, and why I’m looking at it today. The mechanical triage flagged a break in the claims resting on operating margin. The cause is brutal and entirely exogenous: in Q2 2026, the average fuel price Copa paid rose 84.8% year over year, to …/gallon from … Operating margin fell from 21.7% to 8.7% (−13.1 points), net profit from … to … (…), EPS from … to … It isn’t a company problem: revenue grew 25.7%, RASM …, and unit cost ex-fuel fell 0.1%. It’s the global jet-kerosene refining-spread shock — the crack spread hit …/barrel in March 2026 versus a historical ~… and the Gulf Coast spot jumped from …/gallon in January to … in April (EIA, monthly series).

The thesis. Copa is a rare-quality operating asset in its sector, trading at a price that assumes the fuel shock is permanent. Its cost structure (5.7 cents ex-fuel), the Panama hub’s unreplicable geographic position, and a balance sheet with … of liquidity (39% of trailing-12-month revenue) let it ride out the cycle without dilution and without cutting its fleet plan. Refining spreads are among the most mean-reverting series in energy; when they revert toward …-30/barrel, Copa’s normalized earnings power is on the order of …-21/share versus ~… in 2026. You’re buying a cyclical company in a cyclical trough, at 10× the bad year’s earnings and 6.5× normalized earnings.

The counter-thesis, equally serious. You aren’t buying cheap in historical terms. Between 2022 and 2025, Copa traded at an average annual P/E of 5.7-7.3× and EV/EBITDA of 3.5-4.4×; today it’s at 8.7× current earnings and 5.5× LTM EBITDA. The re-rating has already happened (the average annual price rose from … in 2022 to … in 2026). Moreover, the company’s 2026 guidance — 17… operating margin — assumes fuel at …/gallon for the full year; in H1 the average was … and in Q2 … so H2 would need to come in around ~… The EIA spot for August 2026 was …/gallon, i.e. ~…-4.35 all-in for Copa after the into-plane margin. Guidance will most likely be missed, and this is the verifiable falsifier of my short-term thesis (see pre-mortem).

Estimated value. Five triangulated models give an extremely wide range, from … (fuel stays structurally above …) to … (full normalization of the crack spread). Model median: … MOS, i.e. ~…/share. Monte Carlo over 20,000 scenarios, seeded from owner earnings of …: median intrinsic value …, median MOS ** …**, probability of undervaluation …%. The 15-analyst consensus is … — practically identical to the simulation median, which is a useful external confirmation, but also a sign I haven’t found anything the market doesn’t already see.

Verdict: BUY, with a reduced position scaled over time. Business quality justifies holding; the width of the valuation range (P10 at …, P90 at …) and the verdict’s dependence on a single uncontrollable exogenous variable (jet fuel price) call for a half-normal-conviction position, with the second tranche conditioned on the November 18, 2026 (Q3) reporting, which will show whether unit fuel cost has begun to decline. Governance — non-voting Class A, CIASA with 26.6% of capital and 100% of votes, a CEO who since July 2025 is also board chairman — is a permanent reason for a discount, not a new problem.