April 17, 2011
The puzzling differential between the price of oil in different markets seems to be persisting.
For most of the last decade, there was very little difference between the price of West Texas Intermediate traded in Cushing, Oklahoma and that for North Sea Brent in Europe. But a $10-$15 spread between the two developed at the end of January and has remained ever since.
The new gap is essentially a geographic difference between the price paid for oil in the central United States and that paid on the U.S. coasts and anywhere else in the world. For example, Chevron is currently offering $123.25 for a barrel of Louisiana light sweet, $17 more than it is willing to pay for Oklahoma sweet. A year ago the differential was only $3. That gap means that U.S. refiners on the coast are paying a huge premium to buy imported oil, when there are plenty of inland domestic producers who'd love to sell it to them at a significantly lower price.
Gail Tverberg noted that the lower price at the Oklahoma hub resulted in part from pressure of new supplies from North Dakota and Canada. But there's still a deep puzzle of where the violation of the Law of One Price is coming from-- why are producers selling the product in Oklahoma when there's such a much better price to be obtained at the Gulf?
The cheapest way to transport oil from Cushing to Chevron's or somebody else's refinery in the Gulf of Mexico is by pipeline. The Seaway pipeline has the capacity to transport 430,000 barrel a day between the Gulf and Cushing. But currently, the pipeline is carrying oil from the Gulf, where it is so expensive, to Cushing, where it is cheap. Updating Craig Pirrong's welfare triangle calculations, if we assume a pipeline transportation cost of about $1/barrel and that reversing the pipeline at full capacity would be just enough to eliminate the spread, running the pipeline in the reverse direction would generate a combined surplus for oil producers and consumers on the order of (1/2)(17 - 1)(430,000) = $3.4 million every day. If, on the other hand, the added flow was still not enough to reduce the spread, the gain in surplus could be up to twice as big-- $6.8 million per day.
So why does ConocoPhillips, part owner of Seaway, say that using the pipeline to transport crude from Cushing to the Gulf is not in its interests? At its recent analysts call, the company offered this explanation:
the issue there is we have a mid-continent refining center in Ponca City [Oklahoma], and we also want crudes that allow us to make what we call "premium coke" at Ponca. So if there's a need for us to bring crudes into the mid-continent, the other piece on it, everyone talks about reversing Seaway as being a very quick solution. And I would tell you that it's not, you can't do it overnight. The timeframe could be six months, it could be a year. The dollars are not free either, it costs money to be able to reverse it.
reversal of the smaller but longer Spearhead pipeline cost $20 million. The reversal of Line 9 in Canada cost $100 million.
Pirrong concludes that any potential benefits to ConocoPhillips are smaller than the potential gain to other market participants from reversing the pipeline. That leaves room for a consortium of upstream oil producers to profit by offering to buy the pipeline outright, or alternatively to pay a sum to ConocoPhillips to persuade it to reverse the flow, in the spirit of Professor Coase's theorem.
An alternative is a new pipeline extension proposed by TransCanada that would carry oil from Canadian oil sands all the way to the Gulf. That's another option that would clearly cut the Brent-WTI spread, but which has yet to receive U.S. approval.
If you can't ship the product by pipeline, the next best alternative is rail. It should be possible to get oil all the way from North Dakota to a Gulf refinery for $7 a barrel, leaving a very healthy profit for each barrel you ship. The problem here appears to be the infrastructure of rail tanker cars and loading facilities necessary to handle the volume.
But plenty of people are working hard to fix that. U.S. Development Group opened a new 60,000 b/d crude-by-rail unit train terminal in St. James, Louisiana for this purpose last summer, and plans to double the capacity and build two more. Savage Companies and Kansas City Southern plan a huge new terminal for Port Arthur, Texas for completion in 2012:Q2. A dozen other rail facilities for transporting oil from North Dakota and surrounding areas are also under construction. Jim Brown speculates that this may have been part of what Warren Buffett saw that others didn't when he decided to buy Burlington Northern railroad last year.
And even transportation by truck may be profitable at current spreads. If I had any physical assets in this business, I'd be looking into every way imaginable to try to sell North Dakota oil in the higher-priced markets, for the good of North Dakota and for the good of America, not to mention for the good of my own profits. Since I'm only a pixel-pushing college professor, I'll instead just offer a prediction-- arbitrage is eventually going to succeed in driving the Brent-WTI spread down, and nobody-- not even ConocoPhillips, not even the U.S. president-- can prevent it.
But they do have the power to slow it down.
Posted by James Hamilton at April 17, 2011 07:09 AMdigg this | reddit
JDH The BNSF rail lines that run from North Dakota to Louisiana generally follow major rivers, including the Mississippi.
Barge traffic along the Mississippi has been declining and is characterized by an oligopolistic market structure:
The alternative to moving grain down to ports in New Orleans has been via rail. But now we're asking those same rail lines to carry more North Dakota oil down to ports in New Orleans along with North Dakota red wheat. Something's gotta give. The article you cited said oil could be shipped by rail "for as little" as $7 per barrel. That sounds to me like the qualification I might hear from a car salesman. I'm assuming the marginal cost curve is upward sloping, plus someone would have to cover the fixed cost investment, so I'm a bit skeptical of the $7 per barrel figure. I'm not saying that shipping oil by rail can't happen, only that I doubt it can happen at $7 per barrel and without impacting commodity prices for cereal grains.
Posted by: 2slugbaits at April 17, 2011 08:19 AM
The new new deal. Build some fracking freight rail. (Pun intended.)
Posted by: aaron at April 17, 2011 09:43 AM
Are you sure WTI and Brent are identical goods? I heard an analyst say that Brent was sweeter so it requires less refining. If true, then you need to account for the cost of extra refining.
Posted by: tj at April 17, 2011 10:55 AM
tj: On average over 2001-2009, WTI was $1.50/barrel more expensive than Brent.
Posted by: JDH at April 17, 2011 11:14 AM
JDH Could you elaborate on your reply to tj? If Brent is sweeter and requires less refining, then wouldn't that make it more valuable to refiners? And if it's more valuable, shouldn't that bid up the price for Brent relative to WTI? According to your post here:
sweet oil cost more than sour oil because refining costs were lower for sweet oil. Shouldn't the same logic apply within grades of sweet oil?
Posted by: 2slugbaits at April 17, 2011 11:50 AM
2slugbaits: A $10-15 premium of Brent over WTI is a great anomaly, which is why I'm writing about it here.
Posted by: JDH at April 17, 2011 12:30 PM
One word, gentlemen: China.
China is Saudi Arabia's largest customer, recently supplanting the Anglo-American imperial corporate-state.
China's oil consumption and oil imports are growing at 8-10%/yr., whereas Saudi Arabia is at or near peak oil production and oil exports.
At global peak oil production and falling net energy, there is no sustainable surplus global oil capacity going forward. China-Asia's demand for oil and oil exports from exporting countries is the marginal demand effect resulting in a global zero-sum situation for oil.
Hereafter, Peak Oil and population/ecological overshoot means, simply, that real per capita private GDP is no longer possible.
Throw out the models and economics textbooks, fire the e-CON-omists (political propagandists), and shut down the e-CON depts and business schools. We need a new religion to replace the belief in the false god of perpetual growth of population and consumption on a finite planet; it ain't gonna happen.
Posted by: Nemesis at April 17, 2011 12:52 PM
The answer that ConocoPhillips gave was that they wanted to produce premium coke. Which, IIRC, is one of those materials used for those new low maintenance gravel roads.
Also refineries need to be set up for particular mixes of crude. Different grades can be mixed to meet a refinery's configuration, but if that's not an option the refinery needs to be tuned for the new input.
Posted by: aaron at April 17, 2011 01:23 PM
(probable wrong about the roads. I think that was actually coke ash.)
Posted by: aaron at April 17, 2011 01:52 PM
JDH Understand, but in your reply to tj you said that over the long run WTI ran about $1.50/bl more than Brent. I was just wondering if that's what you meant to say or if it was a typo (I can relate to that!). Other things equal I would have expected Brent to run a few dollars a barrel over WTI. Agree that the current differential is hard to explain and that someone is leaving money on the table.
Posted by: 2slugbaits at April 17, 2011 02:00 PM
One factor partly at work is apparently the oil from Canada is not WTI grade, and ome suppliers are diluting WTI with some of this inferior grade, so what is being sold at Cushing as WTI is not as good as traditional WTI.
According to a story on Bloomberg last week, some refiners are buying directly from suppliers to make sure they are getting 100% WTI. Presumably, they are paying some premium to the quoted WTI price.
But it wasn't clear if this accounts for $5 of the differential or $.50. I guess if you knew what the supplier specific contracts paid you could figure it out.
Instead of moving oil around North America, wouldn't it just make more sense to divert some of the East African oil that comes here?
Posted by: Bob_in_MA at April 17, 2011 02:17 PM
2slugbaits: No typo. Traditionally WTI had less sulfur than Brent and sold for a slightly higher price.
Posted by: JDH at April 17, 2011 02:54 PM
Okay, thanks. That makes sense. Like tj, I was under the (apparently wrong) assumption that Brent had less sulphur.
That makes the current spread even more remarkable.
Posted by: 2slugbaits at April 17, 2011 03:29 PM
The futures markets also shows a differential of about $13. WTI can be found here and Brent can be found here. If the simple pipe reversal or alternative transportation schemes were viable in the near term then arbitragers should be coming in on a convergence trade, driving the futures prices closer together in anticipation of those plans being put into play.
Posted by: Tom at April 17, 2011 04:07 PM
Here's the Bloomberg story on WTI quality:
Nymex May Tighten WTI Rules as Refiners Question Quality
Posted by: Bob_in_MA at April 17, 2011 05:00 PM
WTI is both slightly lighter and sweeter, both features that make this spread even more unusual.
Posted by: Tristan Bruno at April 17, 2011 05:23 PM
It's hard to believe coke is big factor in price difference, though it's plausible that there is high local demand for premium coke for specialty industrial applications.
Posted by: aaron at April 17, 2011 05:49 PM
A trucker can haul about 250 barrels of oil in a tanker from Tulsa to Houston or New Orleans for an easy $4000 arbitrage for a day's work. With operational costs of roughly $1.50 per mile, and a round trip, that still leaves a nice net profit, say, $1000 a day. Why doesn't this happen?
Posted by: Anonymous at April 17, 2011 06:05 PM
It's worth noting that the gasoline market seems to be pricing off the Brent market rather than the domestic WTI price. So when economists and pundits refer to oil prices generally, they should keep in mind that the oil price really is already over $120/bbl.
Posted by: tinbox at April 17, 2011 06:08 PM
The continuing divergence doesn't seem that improbable to me (albeit, I'm not an expert in this area):
From Jan to Dec last year, on a monthly basis, Western Canadian output increased by 230,000 bpd, call it 20,000 bpd/month.
From March to November of last year, North Dakota output increased a bit over 100,000 bpd, call it 12,000 bpd / month.
Everyone and their brother wants a piece of the action up there, so it's not hard to imagine volumes might be going up pretty quickly, and that's probably contributing to the bottleneck at Cushing.
As a point of reference, Haynesville shale gas production went from nothing to 2 Tcf per year in three years, according to Platts. Two Tcf is one twelth of US consumption (half of China's)--from one basin!
Could something similar happen to shale oils as well? Don't know. Maybe.
Posted by: Steven Kopits at April 17, 2011 06:10 PM
For my money, a more interesting spread is Brent to Maya sour.
The spread is once again wide. Why?
Is it a lack of refining capacity? (Still?) I had a nice chat with a woman from Valero, and she said their refineries can run both--it's just a matter of which is cheaper.
On the other hand, Brent is widely traded as a financial instrument, Maya is not. Perhaps that's a clue, also an indicator of what's going on in the market.
Posted by: Steven Kopits at April 17, 2011 07:19 PM
Peak oil alarmists are merely the flat-earthers of the modern era. They actually think prosperity levels will regress back to 19th-century levels due to 'peak oil'.
Oil prices will be lower by 2017, than they are today. Oil consumption will be a smaller percentage of world GDP than it is today.
Posted by: GK at April 17, 2011 08:19 PM
Steve, would you be able to find out if the sour is more expensive to.refine, if there are any process changes or partial shutdown necessary to switch, and what the differences in outputs are.
Also I would expect more maintenance for the sour as various filters and catalyst need to be replaced and cleaned.
Posted by: aaron at April 17, 2011 09:18 PM
The WTI-Brent spread has less to do with sulphur, and more to do with the way oil flows. WTI always used to trade above Brent because the US is a big crude importer, and has to pay a premium to secure North Sea oil (and west African, and North African, and Caspian, etc). Benchmarks reflect this. WTI's current discount to Brent complicates this, but doesn't make it impossible. Some US refiners are now paying differentials to Brent for imported cargoes, instead of the usual WTI-based price.
Gasoline doesn't price off Brent, it prices off gasoline. Physical markets trade at a differential to the Nymex Rbob futures contract.
Brent-Maya is getting wider because the loss of Libya's 1.3mn b/d crude exports has triggered a light sweet crude supply squeeze.
Posted by: Down with this sort of thing at April 18, 2011 02:17 AM
It is also likely the spear is price level dependant. As the price level reaches a point, it become a factor in transport and also increases with price.
This does not affect JDHs points.
Posted by: aaron at April 18, 2011 06:00 AM
Demanding function with two variables the price differential between the Wti and the Ns Brent in Europe,the transportation cost where USA coastal refineries are facing an upwards slopes, when it comes to oil flows coming from the USA mainland, to the west coast.
Let us see the recidivists,that is the money managers (P43 P44 IEA report),they are long everything on the WTI (option, futures) and they "like"the WTI light sweet crude oil. The producers are short of the same (may be square against their production contract) Let us face it,judging by the numbers of contracts on the exchange trades, the money managers need oil to make a living,and this much more than the oil producers.
Interesting study, from the IEA economists, when "pealing" the onion future act (P44).After banning the future contracts on onion, the volatility remained the same,and their prices remained still correlated to exchange traded commodities (it may be interesting to breakdown the correlation by items,milk for instance).
So we have seen, an appetite for the light sweet crude oil and the NS Brent is meeting with the flavor,light and sweet it is.What are the countries able to deliver the oil meeting with the definition low API and low sulphur? (IEA report P22)
Algeria ,North sea,West Africa,Caspian Azeri.No mention is made of Irak and yet they are reports granting "Iraq's oil is generally of high quality because it has attractive chemical properties, notably high carbon content, lightness and low sulfur content, that makes it especially suitable for refining into the high-value products. For these reasons, Iraqi oil commands a premium on the world market"
It would be educative to read the balance sheet of the trade bank of Irak and its contingent liabilities.
Aug. 29 (Bloomberg) -- J.P. Morgan Chase & Co. has been selected to operate a bank the U.S. is creating in Iraq to manage billions of dollars to finance imports and exports.
Noted that the detrending prices, Ns Brent WTI occurs before the Lybian turmoils.
Posted by: ppcm at April 18, 2011 07:00 AM
I'm not an expert in downstream. If I have a chance to look at sweet vs sour processing costs, I'll post it.
Posted by: Steven Kopits at April 18, 2011 07:51 AM
Starr Spencer has a great piece on Bakken production over at Platt's Barrel:
Posted by: Steven Kopits at April 18, 2011 08:05 AM
Re: Peak oil alarmists are merely the flat-earthers of the modern era
An interesting metaphor. Flat-earthers held a belief that turned out to be contradicted by reality.
Today, oil cornucopians believe that the sum of the output of discrete regions, e.g., Texas & the North Sea, that show clearly defined production peaks will result in a virtually infinite rate of increase in oil production.
Peak Oilers believe that if regions peak, then the world--the sum of the output of discrete regional producing areas--will show a similar peak.
So, which group is more in tune with reality, especially considering the fact that global crude oil production has been at or below the 2005 annual rate for five straight years?
The 1972 Texas production peak lined up with the 1999 North Sea peak:
Posted by: Jeffrey J. Brown at April 18, 2011 09:28 AM
Re: Shale Plays
I would advise a careful analysis of Art Berman's work on the shale plays.
Posted by: Jeffrey J. Brown at April 18, 2011 09:30 AM
Re: China, et al
Peak oil versus peak exports
Egypt, a classic case of rapid net-export decline and a look at global net exports
From Peak Oil Vs. Peak Exports:
We looked at some near term scenarios for global net exports, out to 2015. We constructed two scenarios. For both scenarios, we assumed a slight 2005-2015 production decline of 5% (0.5%/year) among the top 33 net oil exporters, and we assumed that Chindia’s 2005 to 2009 rate of increase in net oil imports continued out to 2015. The only variable was consumption in the top 33 net oil exporting countries.
For Scenario #1, we assumed no increase in consumption among the exporting countries, from 2009 to 2015. For Scenario #2, we assumed that the exporting countries’ 2005 to 2009 rate of increase in consumption continued out to 2015.
Under Scenario #1, global net oil exports in 2015 would be down by 9.6% from the 2005 level, while the volume of “available” net oil exports, i.e., the volume of net exports not consumed by Chindia, declined by 28% from 2005 to 2015, from 40.8 mbpd (million barrels per day) to 29.5 mbpd.
Under Scenario #2, global net oil exports in 2015 would be down by 14% from the 2005 level, while the volume of “available” net oil exports, i.e., the volume of net exports not consumed by Chindia, declined by 33% from 2005 to 2015, from 40.8 mbpd to 27.4 mbpd.
To summarize Scenario #2, if we extrapolate the 2005 to 2009 rate of increase in consumption by the exporting countries out to 2015 and if we extrapolate Chindia’s 2005 to 2009 rate of increase in net oil imports out to 2015, and if we assume a slight production decline among the exporting countries (0.5%/year from 2005 to 2015), then for every three barrels of oil that non-Chindia countries (net) imported in 2005, they would have to make do with two barrels of oil in 2015.
Posted by: Jeffrey J. Brown at April 18, 2011 09:48 AM
aaron: Premium coke is the essentially pure carbon residual left after everything even remote a hydrocarbon has been cooked off. Much easier to deal with than coal-based coke: essentially zero ash, sulfur already removed, etc.
jh: Unless there are large numbers of suitable tanker cars sitting idle, the marginal cost of rail transport would have to go up quickly as volume increased, wouldn't it? The railroads don't seem to have large amounts of spare rolling stock; a few years ago, bumper crops of wheat sat in piles at the loading terminals in Colorado because the railroads didn't have cars available to haul it.
Posted by: Anonymous at April 18, 2011 09:50 AM
Once again you hit a home run. This is one of the most informative threads I have read in a long time. Thanks to all. The dialogue and discussion has been top level analysis.
Posted by: Ricardo at April 18, 2011 12:02 PM
Why isn't pipeline operation considered by the FTC as a natural monopoly and subject to compulsory separation from production and refining? Isn't this a very similar case to electric power transmission?
Posted by: pdupont at April 18, 2011 12:39 PM
Different markets with different price elasticities due to different opportunities for substitution. A sophisticated retailer prices the same good at different prices in different places and consumers are free to bear any incremental costs associated with shopping about.
Posted by: Frank in midtown at April 18, 2011 01:00 PM
Why do you think the US is fighting a war in Libya for the Frenchies?
Posted by: Jay at April 18, 2011 01:35 PM
GK, oil consumption will be A LOT less because real per capita GDP will steadily decline inexorably hereafter with falling net energy, prohibiting us from consuming at the trend rate of the peak Oil Age era since the 1920s-40s.
Note that the Bekken recoverable estimate is just 3-3 1/2 years' worth of US consumption, albeit about 54-55 years' worth of production at the suggested daily rate.
But the entire recoverable reserves won't come close to making up for the ongoing per capita decline in US crude production since 1970 (45% to date) and 1985 (50-55% to date).
And this does not include the necessary liquid net fossil fuel costs to try to build out wind, solar, wave/tidal, and electric car infrastructure with global peak oil and oil imports resulting in a zero-sum supply situation, as Jeffrey suggests, with China growing oil consumption and imports 8-10%/yr.
As it costs more in liquid fossil fuels to extract costlier reserves, the net energy returns per capita will collapse as soon as '15-'16 to no later than the early '20s.
The net energy costs for Bekken at $106 oil today is a lot less than what we will see for production and flows at $200-$350 oil in the next 10-20 years.
Therefore, it is highly unlikely that we will have anywhere close to the real per capita GDP growth and associated private investment and profits necessary to continue to built out Bekken, fracking, and "alternatives".
Posted by: Nemesis at April 18, 2011 01:47 PM
On NPR's Marketplace:
I gave a long interview to NPR a couple of weeks ago on peak oil issues, of which what survived was the equivalent of "Yeah, what he said." But it was fun.
Posted by: Steven Kopits at April 18, 2011 06:52 PM
What puzzling here is the first sentence:
"The puzzling differential between the price of oil in different markets seems to be persisting."
Then JDH explains the differential in the third paragraph.
The only thing slightly puzzling is why the anti-trust laws do not forbid Chevron from owning the pipeline. But, I guess, that's a pretty naive question.
Posted by: Marcus at April 18, 2011 06:59 PM
The big spread is caused by the cheap US dollar. Euro's want more dollars, where dollars buy them less. American's live in a dollar dominated economy, and could care less what the Euro's think of the value of their dollar.
Trucks cannot be used to haul crude any distance. Canadian crude can be loaded onto ocean tankers at Lake Superior ports, and Oklahoma crude can be shipped from Tulsa to Houston in inland river barges.
Posted by: Joseph Linck at April 18, 2011 07:09 PM
At the end of the day, it will be interesting to see the actual cost of purchased oil in the USA and in Europe. I would guess that current Brent Crude spot and future prices are close to actual prices paid.
So JDH, what do applied economists do going forward? Splice Brent Crude prices into time series of WTI?
I'd be curious to know at what oil price threshold the major European economies risk recession.
Posted by: westslope at April 29, 2011 04:45 AM
It is very difficult to develop an adequate supply response by rail and truck. For rail, it is not just a matter of available rail lines, but one also needs tank cars. The demand to build and lease new tank cars is expanding, but it will take a while for there to be an adequate supply of tank cars to transport enough crude to fix the arbitrage. Same thing for trucks. I doubt there is a substantial number of idle crude carrying trucks sitting in Tulsa. But, if you could find idle trucks and idle drivers, yes, I imagine you could earn a nice arbitrage. But, you had better earn very high returns, because we know that demand for crude and gasoline is declining. Therefore, once this arbitrage is over, there very well may be an oversupply of tanker railcars and trucks.
The arbitrage eventually will be corrected but it will take a long time, perhaps three years or so. Enbridge has proposed reversing certain pipelines in order to move crude west to east rather than east to west, but environmental groups have stalled those plans, just as environmental groups have stalled pipelines moving crude from Cushing to the Gulf Coast.
As to using anti-trust laws to prohibit oil producers from controlling pipeline flows, I suspect that would backfire in the long run, because producers are the ones most interested in spending capital to develop the pipes in the first instance. If we prevent them from controllilng the flows, we likely will indirectly increase the cost of capital to build the pipes in the first instance.
In any event, others, including Chevron's partner in Seaway (Enterprise Partners), have offered to build new pipes moving crude south from Cushing to the Gulf. Which proves that the market will correct itself, albeit perhaps over years, not weeks.
Posted by: Bill Davis at May 22, 2011 01:01 PM