Each week, we comment on the U.S. Department of Energy reports of crude oil and fuel inventories (see our last commentary, "Got Gas? There's Less Now"). As part of our commentary, we depict potential profit margins that oil refiners can obtain by "cracking" crude oil into its major tradable distillates: gasoline and heating oil. This so-called "crack spread" represents the dollars earned, above the cost of crude oil inputs, by selling wholesale lots of refined products.
We get lots of questions about our crack spread computations. Traders and market observers wonder, at times, if we've been breathing cracking tower fumes a bit too long. Allow us to set the record straight.
First of all, we derive our numbers from that most transparent of petroleum marketplaces-futures. As such, the crack spread is generic. It doesn't represent the profit margin earned by all refiners. Or any one refiner, in fact. The crack spread, and the resulting refining margin, is more a rough topographic map than a GPS tracker of profit. Still, even crudely derived - ahem - margins can point to rough or smooth roads ahead for refiners and investors.
For example, some refiners use light, sweet grades of crude deliverable against New York Mercantile Exchange [NYMEX] futures. The NYMEX basis grade is West Texas Intermediate [WTI], a crude that yields more gasoline than more viscous and sour grades. Our numbers are based upon a classic WTI crude refining model (3-2-1) that yields two barrels of gasoline and one barrel of heating oil for every three barrels of oil input.
Other crudes, such as OPEC basket grades, yield less gasoline, so a 2-1-1 (two barrels of crude producing one barrel, each, of gasoline and heating oil) may better describe operations based upon these inputs. Often, news reports referencing crack spreads are based upon 2-1-1 pricing.
Calculating The Spread
Calculating a crack spread requires you to first rationalize crude oil and distillate prices. Crude oil is priced in dollars per barrels, but gasoline and heating oil prices are denominated in gallons. Let's suppose the nearby NYMEX crude futures settled at $40.00 per barrel. That's the input. The refining output is represented by gasoline and heating oil contracts deliverable in the month following the crude delivery. Hedgers and traders use distillate prices a month out from the crude expiration to simulate a storage, refining and marketing cycle. Let's suppose gasoline last settled at $1.28, while heating oil did so at $1.16.
A crude oil futures contract calls for delivery of 1,000 barrels. So too, do the distillate contracts, albeit indirectly. Heating oil and gasoline contracts specify delivery of 42,000 gallons, but with a barrel holding 42 U.S. gallons, it's really 1,000 barrels. Just multiply the distillate prices by 42 to get the barrel prices. Your gasoline, then, fetches $53.76 a barrel, and a barrel of heating fuel, $48.72.
The 3-2-1 crack spread is found through simple arithmetic as:
Gasoline Heating Oil Crude Oil
[(2 x $53.76) + (1 x $48.72)] - (3 x $40.00) = $36.24 per 3 barrels, or $12.08 per barrel, of crude
Similar math is employed to derive a 2-1-1 crack:
Gasoline Heating Oil Crude Oil
[(1x $53.76) + (1x $48.72)] - (2 x $40.00) = $22.48 per 2 barrels, or $11.24 per barrel, of crude
As you can see, there's a difference in crude yields depending upon the refining model employed. Crack spreads are seasonal. The 3-2-1 spread, double-weighted in gasoline, tends to outperform the 2-1-1 spread when gasoline prices rise in relation to heating oil. Seasonally, that's typically in winter and spring ahead of the peak summer driving season. Slackening demand for petrol, on the other hand, can push the 3-2-1 spread to a discount. In the terminal stages of the oil price spiral, for example, gasoline was well-supplied, pressuring the spread under the 2-1-1 crack. Then, at the oil price peak, demand for gasoline dissipated, further deepening the discount. Only since February has the 3-2-1 spread returned to premium and, if history is any guide, points to a May peak.
Crack Spread Models
The ETF Spread
Not everybody, of course, cares to trade in futures. We've mentioned the recent introduction of exchange-traded funds that make a margin-free alternative possible. Think of these collectively as a "good news, bad news" version of the crack spread. The good news is that these ETFs can create a 2-1-1 replicant. The bad news? No 3-2-1 simulant.
Buying equal lots of the ProShares UltraShort DJ-AIG Crude Oil ETF (NYSE Arca: SCO), the United States Gasoline Fund (NYSE Arca: UGA) and the United States Heating Oil Fund (NYSE Arca: UHN) puts you short two units of oil and long one unit each of gasoline and heating oil-à la the 2-1-1 model.
The price of the ProShares SCO fund represents the cost of acquiring two units of oil, while the value of the UGA and UHN funds stand in as the proceeds derived from selling the distillates. Since the introduction of the ProShares SCO last November, the ETF proxy's produced a 13.8% gain versus the 15.1% return generated by the futures model.
2-1-1 Crack Spread (Futures Vs. ETF)
You can see the ETF spread doesn't exactly track its futures analogue. There are two reasons for the disparity. First, the futures spread is based upon settlement - not last sale - values. The ETF spread, however, reflects the last sale data available to retail investors. Using end-of-day bids or NAVs narrows the apparent difference significantly. The difference between the ETFS' last reported sale and end-of-day indicative values has been as wide as 77 basis points (0.77%) recently.
There's another source of dissonance that can't be adjusted away through selective quotation, though. Contango or backwardation isn't reflected in the ETF spread. The product ETFs, by design, are continuously invested in front-month futures, rather than second-month contracts. The current difference between first- and second-month futures runs between a 15 basis point discount for gasoline to an 89 basis point premium for heating oil.
Hopefully, this discourse has cleared the air. Or, at least waved away some of those nagging cracking tower fumes.