Refined products · 6 min read
Reading the 3-2-1 crack spread
The crack spread is a refiner's gross margin. When it widens, refiners run flat-out. When it compresses, they cut runs and crude inventories build. It's a leading indicator of where crude demand is going.
The arithmetic
A refinery takes one barrel of crude and converts it into a mix of products: roughly 45-55% gasoline, 25-30% diesel/distillate, 10-15% jet fuel, and the remainder as fuel oil, propane, asphalt, and refining losses. The exact yield depends on the crude grade, the refinery's complexity, and the seasonal product mix the refiner is configured for.
The 3-2-1 crack spread is a simplified gross margin: assume 3 barrels of crude in produce 2 barrels of gasoline and 1 barrel of distillate (heating oil/diesel). Since CME futures price crude per barrel but gasoline (RBOB) and heating oil (HO) per gallon, you have to convert:
3-2-1 Crack ($/bbl) = ( 2 × RBOB × 42 + HO × 42 − 3 × CL ) / 3
Where RBOB and HO are in $/gallon (multiply by 42 to convert to per-barrel), and CL is WTI crude in $/bbl. The result is the gross dollar margin per barrel of crude processed. Note this is "gross" — not "net." It doesn't subtract refining costs (energy, labor, maintenance, catalyst), which typically run $5-12 per barrel depending on the facility. So a 3-2-1 crack of $15 means a competently run refinery is making $3-10 per barrel net.
Why the 3-2-1 ratio specifically
The ratio mimics a typical US refinery's yield on light sweet crude. A more accurate ratio for the Gulf Coast (which processes more heavy crude with a different yield) would be 5-3-2. For the West Coast, 4-3-1. For Europe (which is diesel-heavy), the 1-1-1 single-product cracks for diesel are more relevant than the 3-2-1 composite.
The 3-2-1 remains the standard because it's a clean, widely-quoted number based on liquid CME futures, and any refinery is somewhere within ±30% of its directional movements. If 3-2-1 widens by $4, every refinery is making more money — even if the exact magnitude varies by facility.
Advertisement
Seasonal patterns
The crack spread is structurally seasonal:
- Spring (Mar-May). Refiners switch from winter to summer gasoline blends, which involves planned turnarounds. RBOB tightens. Cracks widen.
- Summer (Jun-Aug). US driving season peaks. RBOB demand is strongest. Cracks usually widest in late summer if inventories are tight.
- Fall (Sep-Nov). Demand softens. Refiners switch back to winter blends. Another round of turnarounds. Cracks vary widely.
- Winter (Dec-Feb). Diesel/heating oil demand peaks in cold snaps. HO component of the crack matters more. Gasoline demand seasonal low.
What makes it widen
- Refinery outages. Hurricane Harvey in 2017 took ~25% of US refining capacity offline. RBOB spiked, crude dropped, crack went vertical. Unplanned outages are the cleanest crack-widening events.
- Strong product demand. Summer driving season, an unusually cold winter, a jet fuel recovery — anything that pulls products faster than refiners can produce them.
- Crude oversupply. When crude is weak relative to products (e.g. WTI dropping on a Cushing build), the input cost of refining falls while output prices hold. Cracks widen mechanically.
- Product export disruption. If a major foreign refining hub (e.g. Russian refining capacity hit by sanctions or strikes) is offline, US products fill the gap at higher prices.
What makes it compress
- Refinery oversupply. When refiners have been running flat-out for months, gasoline and distillate inventories build. Product prices drop relative to crude. Cracks compress.
- Crude price spikes. Geopolitical events spike crude faster than products can pass through. Cracks compress quickly during Middle East tension because Brent and WTI move first.
- Demand destruction. Recession, COVID-style demand collapses, gasoline tax changes, EV adoption acceleration — anything reducing product demand structurally.
How to use it
Crack spreads are the cleanest single read on US refining equity health. Marathon Petroleum (MPC), Valero (VLO), Phillips 66 (PSX), and HollyFrontier/HF Sinclair (DINO) all see their earnings move with the 3-2-1 in real time. The spread is also a leading indicator for crude demand: when refiners are making good margins, they buy more crude, which firms WTI. When margins compress for 4+ weeks, refiners cut runs and crude inventories build — bearish WTI.
Direct trading of the crack itself is possible via the futures complex — long 2 RBOB + 1 HO, short 3 CL — but spread futures and execution friction make this impractical for retail. Most retail traders use the crack as a signal and trade the refining equities or RBOB outright.
Common pitfalls
- Crack ≠ refiner stock price. Refining equities trade on quarterly earnings, not real-time crack. A crack widening today shows up in MPC earnings in 3-6 weeks. Market may anticipate, but the lag is real.
- Single-product cracks matter for specialized refiners. A diesel-heavy refinery (like a hydrocracker complex in Europe) responds to distillate cracks, not gasoline cracks. Looking only at 3-2-1 misses this.
- Forward cracks differ from prompt cracks. The 3-2-1 calculated off front-month futures is the prompt. Looking at the 6-month-forward 3-2-1 tells you what refiners expect — useful for spotting structural shifts.
What HarborSignal shows
The 3-2-1 crack appears on the Crude Oil Intelligence page with a 180-day history chart. It's one of six inputs into the Crude Bull/Bear composite score. Live values are computed from CME-listed CL, RBOB, and HO front-month futures, so the number you see is the same one refining traders are looking at. Combined with EIA gasoline and distillate inventory data on the same page, this gives a complete view of US refining margin and product balance.