Setup

Phillips 66 and Valero both pre-announced Q1 2026 ahead of their 10-Qs. Both reported realized refining margins +33-48% YoY. Both rallied to within 4-13% of 52-week highs. They look like the same trade.

They're not. PSX added 20M barrels to its net short crude/products derivative position during the Q1 commodity spike — from -33M to -53M bbl, a 60% increase. VLO held its net hedge book flat at -6.4M bbl. PSX took $2.6B of gross hedge mark-to-market losses; VLO's net hedge P&L was +$6M. The Apr 6 8-K pre-announcement framing buried the cohort divergence; the Apr 29 10-Q reveals it.

What the filing says

PSX Q1 GAAP Refining was +$208M (vs -$937M Q1 2025), realized margin $10.11/bbl at 95% utilization. The segment looked positive because $2.6B of gross commodity hedge losses flowed through cost of sales rather than the refining line. The 8-K had pre-announced segment Adjusted -$200M to -$400M.

The hedge book grew. -33M → -53M bbl during the same quarter that lost $2.6B mark-to-market. VaR sensitivity: a 10% commodity move = ±$647M income. The book is going the wrong direction at increasing scale.

The balance sheet stretched. Total debt jumped $7.4B to $27.1B against a $17B end-2027 target. Commercial paper drew to its $5B cap (was $200M). Remaining committed capacity: $0.8B. Q1 operating cash flow -$2.3B (vs +$0.2B Q1 2025). Buybacks continued at $269M.

LIFO reserve expanded $3.6B → $10.5B — $6.9B of unrecognized physical inventory gain offsetting the derivative loss in economic (not GAAP) terms.

CPChem ran 94% O&P utilization, better than the "low-90%" pre-announce. Sales volumes -7% (Middle East JV constraint, US plants compensating). DOW recognized $292M Sadara guarantee impairment same period; revolver call "probable" Q2.

VLO Q1 GAAP Refining was +$1,806M (8.5x PSX). Adjusted vs GAAP gap: $24M. PSX's gap: $400-600M. Same Hormuz, different hedge books.

What the market thinks

VLO is +124% TTM vs PSX +78.5% — historical pair alpha is already in the tape. The question is whether the gap widens further on Q2 hedge-book mechanics.

Option-implied probability of -25% drop in 48 days: ≈46% PSX (put delta -0.46), ≈30% VLO. The cohort dispersion is partially priced in put skew. Equity-side pair pricing is closer to flat.

Insider cohort: PSX CFO Mitchell sold $6.6M, Officer Mandell $7.3M, Director Davis $663K — $14.5M three weeks before the Apr 6 8-K. VLO CEO Riggs and officers acquired ≈$30M in January 2026. Form 4s are Tier 1.

Why the gap exists

Three pieces aren't synthesized in single-name coverage:

Hedge magnitude differential. Reading PSX's derivative footnote tells you nothing about VLO's. Reading both side-by-side reveals the cohort outlier. Single-name analyst structure is set up to miss this.

Disclosure asymmetry. PSX is the only refiner in the cohort that pre-announced segment-specific Adjusted income. The convention is PSX-specific, the hedge book that creates the GAAP-vs-Adjusted gap is too. The headline beat reflects management's own pre-positioning, not durable business improvement.

Insider cohort divergence. Form 4s are public but rarely synthesized as a pair. PSX management coordinated $14.5M of selling in March; VLO management coordinated $30M of buying in January. Tier 1 signal in opposite directions on the same trade.

Risks (ranked, thesis-only)

1. Hormuz resolution. PSX hedges reverse $1-2B; VLO takes pure refiner exposure to crack-spread compression. Pair flips. Worldview-implied probability: 12%.
2. PSX unwinds the -53M short proactively. Removes the forward MtM tail. Q2 10-Q discloses.
3. CPChem ME JV recovery. The positive PSX offset to the short leg evaporates. Sadara has no restart timeline; Q-Chem I/II detail not disclosed.
4. VLO Q1 hedge balance was tactical, not structural. If VLO grows net short into Q2 — converging on PSX's posture — the differential collapses.

Catalysts

No forcing function before late June. The trade unfolds through 6-8 weekly data points.

What would change our mind

• PSX 8-K disclosing hedge book reduced to <-40M bbl post-quarter
• Hormuz diplomatic resolution announcement
• PSX rating watch-negative or buyback suspension (amplifies, doesn't reverse)
• WLK Q1 ethylene utilization < 85% (refutes US structural advantage cohort)
• VLO insider cohort turning seller through Q2

Evidence