Why building a massive oil sands refinery would be a bad idea

There are far more important things for Alberta’s government to spend its money on than subsidizing an oil sands refinery


 
  9
(Shutterstock)

(Shutterstock)

On Oct. 6, the Alberta Federation of Labour (AFL) released a new report that purports to show that refining in Alberta is a viable business opportunity—more than viable, in fact—with internal rates of return projected at between 16 and 25 per cent for a new, greenfield refinery mega-complex. If you could promise rates of return like that, anyone would build a refinery, but are rates of return like that possible in today’s Alberta? Perhaps, but they are only going to be earned if the value in our natural resources has been destroyed in advance. That’s hardly value-added.

The push for more refining in Alberta has gained momentum in the past few years, as the so-called “bitumen bubble”—an Alberta manifestation of discounts for crude in much of North America—saw the gap between Alberta bitumen and refined products grow to historic levels. This occurred because the refined products market largely tracked higher global crude prices. As the image below shows, the value of a barrel of bitumen has, at times, been over $60 per barrel lower than the value of the refined products that could be produced from it. Over the past five years, the value of refined products on global markets has averaged about $45 per barrel more than the value of bitumen in Alberta—a margin large enough, were it to be maintained, to put any refinery solidly in the black.

 

Leach1

Figure 1 Implied bitumen value based on Western Canada Select and Edmonton Condensate prices. Refined products value based on Los Angeles Harbour prices, weighted for standard yield from a bitumen refinery, net of typical pipeline transportation costs. Each barrel of bitumen produces a composite barrel of 0.38 barrels of CBOB and 0.25 barrels of CBOB, both gasoline blending components, as well as 0.33 barrels of low-sulfur diesel, with the residual in petroleum coke, which is valued at zero. Source: Bloomberg

The AFL analysis was developed by U.K.-based economist and former Mobil executive Ed Osterwald. (Ironically, the AFL was apparently unable to find a Canadian firm willing to do the work, and one assumes Osterwald was here on some form of temporary work visa.) The results are based on a project cash flow model of a refinery capable of processing 308,000 barrels per day of bitumen. That’s a big project by any measure—it would be the largest refinery in Canada and among the 10 largest in North America.

Osterwald’s report maintains that this refinery could be built in Alberta for about $10 billion—this despite the fact that a similar refinery project, at less than 20 per cent of the size, is projected to cost over $8 billion, and those costs may well increase before that project is done. When questioned on this assumption, Osterwald maintained that his figure was attainable if the builder chose the right partners, and had appropriate government support—what either of those meant wasn’t really clear. That was, unfortunately, the story of the day as interested readers were left with no clear answers as to what assumptions underpinned the analysis, and the report offers little in the way of clues.

 

Leach2

Figure 2 Refinery after tax internal rates of return, given capital cost and net refining margins. Source: Author’s calculations.

Could a new refinery in Alberta earn a 16-25 per cent rate of return? Sure, it could. Refineries are big bets on small spreads—you spend a lot of money up front to capture the difference between the value of your inputs (in this case bitumen) and the value of your outputs, net your refining and maintenance costs. If you spend $10 billion on a 300,000-barrel-per-day refinery, and it operates for 50 years, you’d need an average, net refining margin of between $33 and $50 per barrel to earn that kind of return. This is not unreasonable given recent history—the average price gap has been over $45 per barrel between bitumen and refined products, and refining operating and maintenance costs are generally below $10 per barrel. If price gaps like the last five years are maintained, a refinery on this scale might still be viable if it cost a more reasonable $18 billion to $20 billion to build.

But will these price gaps be maintained, and what does it mean if they are? The margins any refinery in Alberta will earn depend on pipelines and pace of development—if Alberta is connected to global markets and/or if production is managed to match pipeline capacity, diluted bitumen in Alberta will trade at a value similar to Maya crude traded in the U.S. Gulf Coast, net transportation costs.  It will not capture a world light oil price, of course, but its price will come to reflect its value to refiners globally. To give you a sense of how much that matters, Figure 3 below makes all the same calculations for refining margins, but assumes that pricing relationship—diluted bitumen valued in Edmonton at Maya prices, net a $6.50 per barrel pipeline toll.

Leach3

Figure 3 Implied bitumen value based on Maya and Mont Belvieu Natural Gasoline prices. Refined products value based on Los Angeles Harbour prices, weighted for standard yield from a bitumen refinery, net of typical pipeline transportation costs. Each barrel of bitumen produces a composite barrel of 0.38 barrels of CBOB and 0.25 barrels of CBOB, both gasoline blending components, as well as 0.33 barrels of low-sulfur diesel, with the residual in petroleum coke, which is valued at zero. Source: Bloomberg.

This changes the game significantly—using Maya prices to value diluted bitumen, adjusted for transportation, the average refining margin reduces to about $33 per barrel which, once operating and maintenance costs are taken into account, would likely be reduced to about $23 per barrel—right around the break-even margins you’d need to pay off a $10-billion refinery, but nowhere near enough to cover a $20 billion one. Realized refinery margins are typically this small or smaller—in the first quarter of 2014, Valero reported net revenues for its Gulf Coast operations, before taxes, of $6.19 per barrel refined—and so no one is going to jump into a refinery in Alberta unless they can expect discounted bitumen in the long term.

There’s no question that refineries can be viable in Alberta, if conditions lead to a large enough discount on bitumen. The AFL would have you believe that creating incentives for refining here is acting like an owner—capturing the value from the bitumen and creating jobs, even if that comes via a bitumen discount. It’s not. Through discounted bitumen, we would be transferring value from all Albertans and some extraction companies to some Albertans and refinery companies. In terms of jobs, there aren’t many unemployed oil workers hanging out in Alberta, and so extraction and refining industries would compete for many of the same workers, and offer similar job security—you’re not creating jobs, you’re transferring them from one part of the supply chain to another. From a fiscal perspective, profits earned via bitumen extraction are subject to resource royalties which can be as high as 40 per cent over and above corporate taxes while profits earned by refineries are subject only to corporate taxes. From a value-added perspective, an oil sands extraction plant adds far more value to buried oil sands ore than a refinery adds to bitumen. Discounting bitumen to force a substitution between extraction and refining isn’t acting like an owner as all Albertans should demand our government do—it’s acting like a politician. We should direct our politicians not to spend bitumen in ways they would not be prepared to spend money.

The question Albertans should ask is: if we’re going to spend billions of dollars of government money or billions of dollars worth of bitumen, would we like to spend it offering a better rate of return to refineries or on other things we value like health care, education, and infrastructure—areas that incidentally tend to be more labour- and union-intensive than refining? Count me in for the second choice. Refining is not an opportunity for value-added in Alberta, it’s what will happen if we allow value to continue to be destroyed by selling our resources at a discount. That’s not something that should be celebrated as an opportunity—it’s exactly the opposite.


 
Filed under:

Why building a massive oil sands refinery would be a bad idea

  1. So many implicit and explicit assumptions in this analysis. Let me touch on a few.

    1) if Alberta is connected to global markets and/or if production is managed to match pipeline capacity, diluted bitumen in Alberta will trade at a value similar to Maya crude traded in the US Gulf Coast, net transportation costs.

    Yeah, but what if the Canadian public on either the east or west coast of Canada insists, under no circumstances, will it allow dilbit (diluted raw bitumen) to be shipped by pipeline? I’m not sure this is a given that they will.

    2)Osterwald’s report maintains that this refinery could be built in Alberta for about $10 billion—this despite the fact that a similar refinery project, at less than 20 per cent of the size, is projected to cost over $8 billion dollars,

    I presume you are referring to the NWU which includes carbon capture. If so, this would seem like a costly bit to include when comparing capital costs.

    3)Realized refinery margins are typically this small or smaller—in the first quarter of 2014, Valero reported net revenues for its Gulf Coast operations, before taxes, of $6.19 per barrel refined

    I suspect some of these refineries have been around for ages, have been added to/modified numerous times and are configured to handle a range of feedstocks. All of which reduces efficiency. Would a greenfield site utilizing economies of scale and tighter product feedstock specs, designed from scratch have lower operating costs/higher margins? Not a refinery guy, but I would think they would. Maybe significantly, so I’m not sure that is a valid comparison.

    4) I find it confusing when you lump in upgrading (converting raw bitumen into synthetic crude) with refining. There is a possible intermediate step that, in my view, should be analysed separately. There may be a biz case for this partial step rather than going the full monty.

    • ) Osterwald’s own comment: landlocking oil to create a crude discount is stupid: http://www.youtube.com/watch?v=UGmziqY2p6M&feature=youtu.be 7:40. As I said in the piece, if that happens, refineries will do well, but it will be a symptom of foregone value, not value-added.

      2) I presume you are referring to the NWU which includes carbon capture. If so, this would seem like a costly bit to include when comparing capital costs.

      Right, but NWU is looking at capital costs of 160k/bbl/d. The highest costs that I used in my analysis were half that. I talked about a reasonable number being, perhaps, 65-70k/bbl/d of bitumen throughput.

      3)Would a greenfield site utilizing economies of scale and tighter product feedstock specs, designed from scratch have lower operating costs/higher margins? Not a refinery guy, but I would think they would. Maybe significantly, so I’m not sure that is a valid comparison.

      Likely, but all of my analysis is based on lower costs per barrel of 7-10 all-in for maintenance and op.

      4) I find it confusing when you lump in upgrading (converting raw bitumen into synthetic crude) with refining. There is a possible intermediate step that, in my view, should be analysed separately. There may be a biz case for this partial step rather than going the full monty.

      I’m not lumping upgrading in at all, I am talking about refining. Based on the numbers I’ve seen, upgrading to SCO costs pretty close to as much as full conversion refining (see capital costs for Voyageur as an example) and yields a lower-value end product. I think it’s possible, although really tight, to make the case for a refinery, especially if you have pipeline constraints as I say in the piece. Can’t see how you’d make a case for an upgrader.

      • OK, thx. Might be worth revisiting in a few years once the outcomes of key pipelines are finalized.

      • On 1)

        Although it is blended to have a similar API gravity to Maya, their will likely continue to be a material quality discount for Bitumen Blend relative to Maya going forward. It is a “dumbbell” crude meaning it is lean in the middle distillate range that refiners crave. With more expenses involved to address stakeholder concerns for new pipelines and extended timelines I would offer $10/bbl as a more reasonable “equillibrium” transportation. Currently, with rail still in the picture, incremental costs are closer to $20/bbl.

        On 2)

        Costs for NWU include the air separation unit and gasifier that produce a concentrated CO2 effluent just like the subject refinery in the Osterwald study. A separate entity was set up for the Carbon Capture. The refinery in the Osterwald study was taken from the earlier Nestor study and includes petrochemicals as well. It is more complex than NWU. I am not sure the $160,000/bbl of NWU should be taken as gold standard for reasonable, but I would be very surprised if the Nestor/Osterwald refinery could be built in Alberta for less than $150,000/bbl of dry bitumen. Reality check: Complex refining assets in North America can be had for less than $10,000/bbl throughput (see Valero).

        On 3)

        As a big shareholder in Valero I would wager the cost structure of their existing refineries will beat this refinery in Edmonton. Significant refinery cost are associated with production of hydrogen which typically depends on cost of NG and thermal efficiency of units used to produce this hydrogen (Steam Methane Reformers). The Osterwald Refinery instead chooses to gasify pet coke to produce hydrogen. It trades capital cost /sustaining capital for reduced requirements for NG. You typically need much higher NG prices to justify this choice

        On 4)

        Given public metrics for the value of refining assets I believe we will always be better off if we try to mesh with existing refinery infrastructure rather than duplicate it. And I don’t think Voyageur should be considered the poster child for upgrading in Alberta. It was designed to produce a sweet SCO and required hydrogen production / hydrotreating facilities that more than doubled the cost. With low NG prices, sunk capital, and significant refining “swell” through hydrotreating, refiners can afford to pay more for sour feedstock than ever before. I think the sweet spot for value add in the province is full deep conversion to produce sour products. It can be accretive and supportive to upstream investment. It can be done for a fraction of the cost of a Voyageur and it produces at product that should be coveted by any refinery in North America. Market access costs areslashed.

        FWIW.

        • All good points. On dumbell crudes, you’re correct re: dilbit. If you’re comparing to a Maya barrel, the closest assay-wise is likely to do a crude-bit blend with a west texas sour, which is one of the reasons bitumen by rail is gaining traction on the gulf coast. On transportation costs, I tend to agree that the marginal barrel at Edmonton is likely to be more discounted, but most of that nets out if you consider that I was using the same toll to get your refined products out to global markets, and using LA Harbour vs Gulf Coast, which is still a premium market for refined products. On sour synthetics, that’s an interesting thought – I haven’t considered that, but it would certainly be worth looking at. I also think you’re absolutely correct on the value of refining assets – see Citgo for a prime example. If you want to buy a complex refinery, there are some available today.

          Thanks for reading!

  2. Exporting tankers full of diluted bitumen from the West, East and North coasts might seem like a good idea until the inevitable bitumen spill. Then the ensuing ecological disaster will be of such magnitude later generations will wonder what in the hell we were thinking.

    The first dilbit spill already happened in a pipeline rupture in the Kalamazoo River causing “one of the costliest spills, in U.S. history.”

    Some economists like to belittle David Suzuki’s impression of “externalities” as costs businesses pass onto society that economists ignore (“another word for ‘we don’t give a sh*t.'”) But this is a prime example.

    Oil companies are gambling with other people’s future. Not only in bitumen spills, but in the vast amount of carbon being released getting this very dirty, low-energy-return oil to market. This is a huge moral hazard and a huge “we don’t give a sh*t.”

    All right-wing economists who support exporting Alberta’s unprocessed bitumen by tanker to China, India, Europe, or wherever, owe Mr. Suzuki a big apology. Hopefully we will come to our senses before suffering yet anther disaster at the hands of free-market fundamentalists.

Sign in to comment.