Over the past few years, three different Midwest refineries have been quietly working on projects to improve their heavy crude processing capacity. In total, these investments will increase their previously limited coking capacity by nearly 200 mb/d.

The first project, consisting of a 65 mb/d coker and a 50 mb/d crude unit expansion at ConocoPhillips and Cenovus’s Wood River refinery, was completed in November 2011 and began operation in early 2012. By 4Q12, Marathon expects to finish construction on a 28 mb/d coker at the company’s Detroit refinery; that investment also includes a small expansion of atmospheric distillation capacity by 14 mb/d. The largest addition in heavy crude processing capacity will stream in early 2013, when BP’s Whiting facility will bring online a 102 mb/d coker.

Currently, Detroit, Whiting, and Wood River source a significant portion of their crude inputs domestically -- primarily light WTI and WTS type crude. PFC Energy estimates that some 310 mb/d of the 730 mb/d processed at the three facilities in 2011, or 43%, came from US fields. Of that, nearly all was light (or medium-light) crude from the Permian Basin. Much of the remainder, some 350 mb/d, was synthetically upgraded Canadian oil sands crude, supplemented by more limited volumes of heavy, sour diluted bitumen. However, as the coking units are streamed, the crude slates of these three refineries will shift dramatically. The refineries will be obliged to process much greater quantities of heavy, sour crude in order to fill their coking units. PFC Energy projects that essentially all of this incremental demand will be met by imports of diluted bitumen from Canada.

Expected Additional Demand for Canadian crude

Although much of Canadian bitumen production is synthetically upgraded to light, sweet crude, a large fraction is simply blended with condensate or some other diluent, lowering its viscosity enough to be transported via pipeline. This diluted bitumen typically has an API around 20 degrees and sulfur content in excess of 3%, similar to Mexican Maya. However, diluted bitumen also has a relatively high acidity (TAN) and often contains minerals (e.g. nickel, vanadium) that require additional processing to remove. Due in part to the added challenges of refining diluted bitumen, its prices are heavily discounted -- historically trading at a discount of around US$ 5/bbl to Maya and around US$ 15 to 20/bbl less than light, sweet crude. This discount to even other heavy, crude benchmarks underpinned each of these coking investments; the expected future feedstock cost savings (assuming wide light-heavy differentials) far outweighs the significant capital expenditures for these coking projects.

However, to take full advantage of this potential, a refinery must have enough residuum to fill its coking units. Typically, diluted bitumen yields a residuum fraction of greater than 35%. But given the amount of coking capacity being constructed, Detroit, Whiting, and Wood River would collectively need to process over 600 mb/d of diluted bitumen (or an equivalent heavy, sour blend) in order to fill their coking units. For comparison, these three refineries only processed about 90 mb/d of heavy crude in 2011. Thus, the incremental demand for heavy crude in upper PADD 2 could exceed 500 mb/d by 2013. And given that the crude distillation capacity of these three facilities is only being expanded by a collective 64 mb/d, most of this incremental heavy demand will simply displace the lighter barrels previously consumed.

PFC Energy estimates that by early 2013, when all three projects are expected to be fully streamed, Detroit, Wood River, and Whiting’s demand for light crude could be as much as 475 mb/d below the 2011 baseline. Of course, such a massive shift in crude flows will not happen overnight; supply contracts, upstream production, logistical bottlenecks, and market rates will all influence the exact crude slate of these three refineries.

The largest increase in heavy crude demand will come from BP’s Whiting refinery, located just outside of Chicago. This refinery, which has access to Enbridge’s massive Mainline System, could need as much as 260 mb/d of additional heavy crude to fill its soon-to-be-completed coking unit. Marathon’s Detroit refinery, which would need an additional 80 mb/d of heavy crude after its coking project is streamed, is also connected to the Mainline, although further down the pipe. Wood River, whose coking project is already complete, has likely increased its demand for heavy crude by some 195 mb/d against 2011. This refinery, co-owned by Canadian oil sands player Cenovus, has a variety of options for sourcing Canadian crude, including Kinder Morgan’s Express Pipeline (via PADD 4) and TransCanada’s Keystone Pipeline.

Ironically, while each of these projects was commissioned to take advantage of the significant discount for diluted bitumen, the dramatic increase in upper PADD 2 heavy crude demand will place upward pressure on the price of Western Canada Select (WCS) and other such blends. Indeed, these three refineries will have the capacity to absorb all of the projected growth in oil sands production through 2014. Thus, light-heavy differentials in the Mid-Continent will certainly tighten, bringing diluted bitumen pricing more in line with global heavy benchmarks, undermining the margin potential of more sophisticated refineries.

What Happens to the Light Crude?

While the three coking projects are likely to affect heavy crude prices and differentials, their most dramatic impact may be on the price on light inland crudes. As noted, when all three cokers are online, they could displace up to 475 mb/d of the light to medium US and Canadian crude these refineries previously processed. With imports of most non-Canadian crude to PADD 2 already effectively eliminated, this “homeless” crude will be forced to find its way to other refineries, likely those on the US Gulf Coast. This will, in turn, back out Gulf Coast imports of light Atlantic Basin crudes from Algeria, Nigeria, and Saudi Arabia.

However, little to no open pipeline capacity currently exists between points inland and the US Gulf Coast, meaning the displaced light volumes have to be transported via alternative means – rail, barge, and (if the price is right) even truck. The added cost of moving these marginal barrels to the Gulf Coast will put downward pressure on the price of light crude at points inland. Indeed, this dynamic is already underway; with Wood River’s coking unit now online, light barrels previously consumed by this refinery have been replaced by heavier barrels.

As discussed above, the potential volume of light crude being displaced exceeds 150 mb/d. Perhaps because of this very displacement, the WTI-Brent differential, which had stabilized at around US$ 10/bbl since November 2011, has grown to around US$ 20/bbl in recent days. Of course, some of the widening is due to global dynamics boosting the price of Brent, but it is clear that the volume of light US and Canadian crude arriving at Cushing has also grown significantly, depressing the price of WTI.

Crude Dynamics Shift as Pipelines, Cokers Come Online

While the coking projects could ultimately back out as much as 475 mb/d of Mid-Continent light crude inputs, pipeline capacity between points inland and the Gulf Coast is also being increased. The timing and volume of these capacity increases relative to the completion of the coking projects will dictate how wide the WTI-Brent differential grows.

As noted, there is currently no available pipeline capacity from points inland to the Gulf Coast so the light volumes being displaced at Wood River are already pushing down prices at Cushing. By mid-2012, however, the first phase of the Seaway pipeline reversal project should be complete. This pipeline, which links Cushing and the Houston area, will have an initial capacity of 150 mb/d. Thus, Seaway should be able to “offset” most if not all of the light crude displaced by Wood River, thereby tightening the WTI-Brent differential. However, when the Detroit coker is streamed in 4Q12, additional volumes of light crude will be backed out and pipeline takeaway capacity will once again be exceeded.

In early 2013 (the exact month is uncertain at this point), the massive Whiting coker project is expected to be brought on stream, potentially backing out more than 250 mb/d of light inland crude. However, around the same time, Enbridge and Enterprise expect to increase Seaway’s capacity to 400 mb/d, effectively offsetting the impact of Whiting on inland crude balances. The balance should begin to fall only by mid- 2013, when another pipeline reversal takes effect. Magellan expects to complete the reversal of its Longhorn Pipeline, which will connect El Paso with Houston and have a capacity of 135 mb/d, by mid- 2013. Again, the exact timing of these projects relative to one another will have a major impact on balances and, by extension, the WTI-Brent discount.

On top of the aforementioned dynamics, the inland light crude supply balance will also be affected by production growth from the Bakken. After all, if incremental output from this unconventional play is excluded, pipeline capacity from the Mid-Con to the Gulf Coast will catch up to the volumes displaced by the three coking projects by mid-2013. However, the Bakken has consistently exceeded industry expectations and PFC Energy estimates that production from the region could grow by an additional 150 mb/d through 2013, which will drastically impact the price of all inland light crudes. None of the aforementioned pipeline projects do anything to help Bakken crude find a home. Pipeline takeaway capacity in that sub-region is already well below production and is expected to remain as such for the next few years, forcing producers to rely on rail for transport to demand centers.

Looking forward, PFC Energy projects the WTI-Brent differential to average about US$ 18/bbl for 1Q12. As noted, not only is the additional supply of inland light crude depressing the price of WTI, but global dynamics are also helping to boost Brent. Indeed, the spread between these to markers is expected to average US$ 22/bbl in 2Q12. When the first phase of the Seaway reversal opens in July 2012, most of the “stranded” incremental medium to light crude volumes will have a way out of the Mid-Continent. As a result, PFC Energy estimates that the WTI-Brent spread should narrow to around US $15/bbl in 3Q12. That improvement will be short lived, though, since the streaming of Detroit and Whiting’s coking units will back out still more light crude. Thus, the WTI-Brent spread is expected to move back out to the US $18 to US$ 20/bbl range, with no improvement likely until mid-2013 when the expanded Seaway and Longhorn pipelines are completed.

(Rob Smith is Managing Director of PFC Energy's Downstream Group.)