CACI: 34 Points the Cohort Hasn't Priced

CACI International is a federal IT services contractor that just printed Q3 FY26 with organic growth accelerating, margins expanding, and federal civilian revenue growing — while every comparable peer is bleeding. The 10-Q (filed 2026-04-23) closed the $2.64B ARKA acquisition and pushed leverage to 4.6x. The stock fell 11% in the month into and through the print. This memo is about whether the customer-mix divergence between CACI and its cohort is in the price.

What the filing says

Q3 FY26 (quarter ending March 31, 2026):

• Revenue $2,351M, organic growth +6.8% (accelerating from +5.6% YTD)
• Operating margin 9.73%, +67bps YoY — while absorbing $22.1M of ARKA M&A transaction costs
• Operating cash flow +30% YoY ($508M nine months)
• Backlog $33.4B, +6.4% YoY; RPO $12.3B
• Revenue mix: DoD +9.7%, IC +5.3%, Federal Civilian +6.7% (despite DHS partial shutdown since Feb 14), Commercial +19.5%
• ARKA closed March 9 for $2,642.7M (largest acquisition in CACI history); only 22 days of revenue contribution ($28.5M)
• Total debt $5.2B vs $2.9B at fiscal year-end; covenant compliant; $1.07B revolver available; buybacks suspended
• Fourth Circuit affirmed $42M Abu Ghraib verdict 2-1 on March 12; Abbass case (46 plaintiffs) stayed pending SCOTUS Cisco v. Doe

What the market thinks

CACI sold off post-print. RSI 25.5 (oversold). P/E 22.1x vs cohort 11.7-13.6x — a premium that recognizes some differentiation but prices Q4 as if it converges to peer trend, not as if it extends the divergence. Options structure is bullish (OI P/C 0.33, max pain $560 / +8.4%, IV Rank 111%) — institutional positioning leans long even as the tape drifts down.

September 145d ATM IV is 43.8%. Bracketed against our scenarios:

Market-implied figures are derived from option strike deltas at 145d expiry and should be read as approximations, not point estimates. The asymmetry — ≈15-20pt gap on the upside, ≈5-6pt gap on the downside — is the tradeable claim.

Why the gap exists

The market sold the federal services cohort indiscriminately on DOGE fears in late 2025. Customer-mix divergence then appeared in audited GAAP segment data over the past two months but has not been synthesized:

• CACI federal civilian: +6.7% YoY (Q3 FY26, filed Apr 23)
• BAH civil: -27.7% YoY (Q3 FY26, filed Jan 23)
• ICFI federal: -25% YoY (FY25 reported Feb 26)
• SAIC: -3% organic FY26, guides FY27 -2 to -4%

The 34-point spread between CACI fed civ and BAH civil in the same DOGE/shutdown environment is the cleanest single-quarter signal that DoD/IC concentration acts as a structural shield. CACI is 78% DoD+IC. BAH is roughly the inverse. The OBBBA reconciliation funding ($156B defense + $25B Golden Dome + $170B border) is available even during civilian shutdowns — a regime feature the cohort selloff didn't price.

A factor regression versus SPY/ITA/XAR/BAH/LDOS shows CACI clustering with federal services peers, not defense primes (negative loading on the broad defense ETF). Idiosyncratic variance is 55% — below the 75% target — meaning a third of CACI's variance is peer-cohort co-movement. This is why the basket pricing has been able to drag CACI down despite its print: the peer beta is real. The pair structure (long CACI / short BAH) isolates the divergence factor; long-only retains cohort beta as a feature or flaw depending on how the cohort moves.

Honest forward alpha estimate: 8-15% annualized after estimation error, backward contamination, and the partial multiple premium already in the 22x P/E.

Risks (ranked by impact)

1. Leverage 4.6x net debt/EBITDA. Buybacks suspended; capital fully subordinated to debt service for ≈24-36 months to reach 3x. Caps multiple expansion. Requires ≈$500M+ FCF annually; current pace supports it but no room for additional large M&A.
2. ARKA integration opacity. $2.64B acquisition with no standalone revenue, EBITDA, or contract-mix detail disclosed. Q4 FY26 (reports late Jul/Aug) is the first full quarter and the proof point.
3. Abu Ghraib Abbass tail. 46 plaintiffs at the same per-plaintiff rate as Al Shimari = ≈$2.7B potential exposure. Currently $0 accrued. Tied to SCOTUS Cisco v. Doe — not imminent, not in the price.
4. Cohort re-rating risk. If BAH civil decline narrows in May 22 print, the divergence weakens before peer-print confirmation arrives.

Catalysts

What would change our mind

• BAH civil decline narrows to <-15% YoY in May 22 print → divergence priced or normalizing
• LDOS prints below mid-singles organic May 6 → "DoD/IC shield" narrative weakens
• CACI Q4 organic decelerates below 5% or operating margin contracts below 9.5% → ARKA dilution materializing
• Abbass case activates against CACI in active litigation → tail risk becomes near-term
• OBBBA defense funding renegotiated or Golden Dome restructured → forward catalyst removed

Evidence

LR signal: 1.4 — bullish, market underweighting the customer-mix discriminator.