Peak Oil Barrel

I added 2007 to the data, and wanted to create a similar insightful graphs as I saw earlier from Rune Likvern to show the total contribution of past wells to current output. As you can see, recent production is taking a larger % of the current output when time goes by, due to the ever increasing decline rates, where we already saw that 2013 wells decline 60% in the first year this will be even more extreme in 2014.
25 million of the 29 million barrels of crude reported by ND in Nov 2013 is included, as some wells are not reported as individual wells, or missing/incorrect data had me to leave them out. 2007 also contains everything pre-2o07, which is very little anyway, relatively.

Wes538

02/12/2014

These analyses are really excellent, excellent work, Enno. Great job.

Just out of curiosity, what did you use to export the data from the monthly report PDFs into a more usable form? I have tried to do this in the past, but have personally come up short on finding a functional method.
Enno

02/12/2014

Thanks Wes. I initially also thought it would not be so easy, but after I copied one report to a text file, I realized that they are very consistent in structure(same structure from 2007 until now), and not too difficult to parse. After that I downloaded the other reports, copied each to a text file (total half hour work), and used the same parser to read all of them. I didn’t use excel though, as I guess it won’t be able to handle the almost 400.000 well report records, and it’s not easy to add more structure to it.

Don Westlund

http://www.middlefield.com/pdf/resource/012914.pdf

Comments from Henry Groppe about shale oil and Iraq.

Watcher

02/11/2014

That guy is 87.
D C

02/11/2014

Don,

Thanks for that link, very interesting.

Old farmer mac

If I’m eyeballing this graph correctly,and understand what it shows about fifty percent of the total ND production last year came from last years new wells. It’s hard to imagine how anybody who sees this can expect tight oil production to keep rising sharply more than another half decade or so but people will believe anything so long as they want to.

Here’s a link about current Russian plans for exploration and production in the Arctic.
It looks as if Rosneft is going to build in house shipyards for the express purpose of building drilling platforms.
http://oilprice.com/Energy/Energy-General/Russias-Rosneft-to-Build-Arctic-Seagoing-Vessels.html

Jeffrey J. Brown

02/11/2014

As noted up the thread, the average oil production rate for Bakken wells in the first half of 2013 was about 135 bpd. Since, as you noted, the average is distorted* by (temporarily) new high production rate wells, the median rate has to be quite a bit lower–I would guess that the median is in the 60 to 70 bpd range.

And the median production rate would be the killer in higher cost operating areas like Siberia and the Middle East–compounded by the fact that not all shale/tight plays are commercial, and those that are commercial tend to be more gas prone that oil prone.

*Assume 11 wells with the following bpd production:

10, 20, 30, 40, 50–60–70, 80, 90, 100, 1,000

The average production rate would be 141 bpd, but the median production rate would be 60 bpd. In higher cost producing areas, the bulk of these wells might be non-commercial, with maybe two or three wells (or maybe just one well) being a commercial producer. Also, in higher cost areas, they won’t get any long term benefit from the hyperbolic “tail” of production.
Enno

02/11/2014

I looked at the data for the median well in Nov 2013. If you look at all individual wells, that were actually producing (some are permanently, or temporarily not producing) :
average bpd : 141
average barrels per month : 3693 (not all wells produced all days)
median bpd : 93.33

But as you say, the variation is still very extreme, which the average didn’t show at all. The top 10% produced more than 290 bpd, while the bottom 30% produced less than 52.
Doug Leighton

02/12/2014

Jeff,

Do you have any information on, or feeling for, which significant country will be next in making the exporter-to-importer transition? To my mind this is as important a PO metric as any. I expect you’d agree. It’s interesting to monitor how this process is playing out in affected countries; of course there are always mitigating circumstances. Clearly Egypt and Argentina are in disastrous straights with Indonesia not far behind.

From personal observations (was in Hanoi last Fall) I know Vietnam is coping fairly well, at least at the moment. Of course, the VN economy is growing from a low base and people there are especially resourceful. The UK is possibly a more interesting case, at least to me. From the hype, and on the surface, the economy is well into recovery mode. However, even in bonny Scotland, where I also spent some time last year, people, especially in rural areas, were complaining about how high fuel (gas) costs were rapidly eroding the viability of their farm operations. Of course farmers are chronic complainers but similar grievances are common all the way to London.

Anyway, I look forward to your views on these matters.

Doug
D C

02/11/2014

Hi Old Farmer Mac,

It’s a little tricky because you have to look at the area under the curve from Nov 2012 to Oct 2013 (the most recent 12 months). It is probably easiest to just look at a single month such as Oct 2013, and for that month you are correct, about half the output was from well that began producing in 2013. If you look at the nov 2012 to oct 2013 output more than half is from 2012 and 2013 wells, but 2013 wells alone would be about 30 % of total output over the last 12 months shown on that chart.

D C

Hi all,

Enno has shared his Bakken well data and using an Arps hyperbolic to fit the average data from 2010 to 2013,
I get qi=10893 b/month, b=1.4, and Di=0.173 for an average well profile. At 115 months from first output I assume an exponential decline at a 7 % annual rate, the 30 year EUR is 381 kb. Also of interest is Bruno Verwimp’s suggestion that the shape of the well profile (determined by b and Di) is likely changing. I tested this by looking at how the average well profile matches 2008, 2009, and 2010 data when only the qi parameter is changed in the hyperbolic equation. With b=1.4 and Di=0.173 for all 3 years, I find qi is 7692 barrels per month for 2008, 9348 barrels per month for 2009, and 10641 barrels/month for 2010. The match with the data is pretty good as shown in the chart below. The average well profile for 2010-2013 data is the dashed line (which is almost on top of the 2010 line so is hard to see).

Watcher

Anyone want to bet that when the Bakken rolls over some redefinitions take place of what is oil or what is a month or what is production or what is this or that to corrupt the numbers and further render projections suspect?

Watcher

02/11/2014

Speaking of which, this from ZH:

http://www.zerohedge.com/news/2014-02-11/golden-age-gas-possibly-interview-iea

The incompetent IEA chieftess gave an interview and said this:

“OP: Are there any other ways around the “crude wall” aside from lifting the export ban?

IEA: As we wrote in our January 2014 Oil Market Report, much of the LTO is produced in the form of lease condensate, which is most optimally processed in a condensate splitter. There is currently only one such facility in the United States, although at least five others are in various stages of planning and construction.

I mention this issue because one could imagine a scenario under which lease condensate is excluded from the crude export restriction. The US Department of Commerce, which enforces the export ban, includes lease condensates in the definition of crude oil. However, this definition could be changed, or the Commerce Department could simply issue lease condensate export licenses at the behest of the President.”
Watcher

02/11/2014

Previously in the interview she was leaning on — these are matters for the US internally and we don’t comment blah blah blah and then WHAM . . . she just happens to know the agency enforcing the ban and how to redefine words to make it happen.

What a scam.
aws.

02/11/2014

IEA: As we wrote in our January 2014 Oil Market Report, much of the LTO is produced in the form of lease condensate,

Well there you have it!

Nice catch Watcher.
aws.

02/11/2014

Also from the OMR…

the 525 kb/d Northern Gateway Pipeline, received conditional approval in December. The federal government’s Joint Review Panel approved the project, which would carry crude from Alberta to a British Columbia port, subject to 209 conditions and final ratification by the federal government. The pipeline will have a reverse‐flow pipeline constructed beside it in order to import as much as 193 kb/d of diluents.

Funny how Alberta now has to import light hydrocarbon liquids.
D C

02/12/2014

Hi aws,

You realize, I hope, that the condensate is imported to Alberta so they can get the tar sands to flow in pipelines after they mix the bitumen(aka tar sands) with the condensate. This is not anything new or unexpected. It is hard to move landlocked extra heavy oil.
Watcher

02/11/2014

Well, Ron did find a legit Platts article comparing Bakken diesel and kerosene content to Louisiana Sweet’s content (though the original point was a comparison to Nigerian fractions — maybe LS is known to be so and not the right yardstick, dunno.).

So this doesn’t put anything altogether to bed, but it is a damning phrase from her. My focus was on the willingness to redefine things to get a result desired. The graph folks could be setting themselves up for a disaster, but that should not deter anything. Awareness suffices.
D C

02/12/2014

This is not a change in definition for the EIA. For a very long time the US has reported C+C output which means crude plus lease condensate.
Watcher

02/12/2014

You missed the point. The change would be to separate them and define condensate as NON CRUDE to allow export.
D C

02/12/2014

Hi Watcher,

I intended that as a response to AWS and should have been clear about who I was responding to.

I did in fact miss the point as I skimmed through too fast and interpreted aws’s comment backwards.

aws wrote:

“IEA: As we wrote in our January 2014 Oil Market Report, much of the LTO is produced in the form of lease condensate,

Well there you have it!

Nice catch Watcher.”

And earlier you had talked about redefining stuff when the Bakken declines (if I interpreted that correctly), I missed the exports part and how that redefinition was to allow exports of condensate.

The EIA seems to be moving more and more towards the total liquids definition, so stuff is going to be redefined in whatever way the government thinks best.

It would be nice if liquid fuel was reported in Energy terms rather than by volume or mass.

Sorry for not reading more carefully, I did indeed miss the point.
aws.

02/11/2014

From IEA’s January 2014 Oil Market Report … (p.26 sidebar, pdf)

Much of the LTO is produced in the form of lease condensate, which is most optimally processed in a condensate splitter. There is currently only one condensate splitter in the US, although at least five others are in various stages of planning and construction.

— snip —

An additional outlet to expand exports under current legislation would be to exclude lease condensate from the crude oil export restrictions. The US Department of Commerce, which enforces the export ban, includes lease condensates in the definition of crude oil. However, this definition could be changed, or the Commerce
Department could simply issue lease condensate export licences at the behest of the President.

Old farmer mac

IIrc, the first time I encountered this political card trick Nixon was prez and unhappy with the housing start data.

So somebody decided that house trailers are houses and presto, the house starts looked great for the republicans in the upcoming elections.

The first time this sort of thing was actually recorded it was most likely on a clay tablet with a stylus.

If this were a political site I could post a number of current examples of this trick of renaming things for partisan or special interest ends.

I’m inclined to think Watcher is right this time.

In my estimation it doesn’t really matter much either way insofar as our US oil supply is concerned because oil is about as fungible and as easily substituted from one supplier to another as any commodity so long as transportation is available and there is still plenty of it being actively bought and sold on international markets. If it gets to the point that we can’t import enough to make up for any exported oil, then the law against exporting oil will be renewed because that is the way our political game is played at this point in time.

Unfortunately if exports are allowed then the happy go lucky camp will trumpet it to the moon like coyotes as evidence that happy motoring is back and here to stay.This will in turn have the extremely ill effect of setting back energy efficiency and conservation substantially.

The bigger they are the harder they fall. When the inevitable oil crunch finally arrives the folks who championed exports are going to look like the proverbial hooker in church and you won’t hear a peep from them about the virtues of free trade in oil.

aws.

Oglala Sioux vow to stop Keystone XL on the ground if Obama won’t say no

Indigenous group has been touring United States and Canada offering direct action training to nations committed to standing with the Sioux.

Erin Flegg, Vancouver Observer, Posted: Feb 6th, 2014

aws.

02/12/2014

CGA Canadian gas storage levels

islandboy

02/12/2014

LIAT bleeding Barbados dry

In the post, Mr MacLellan highlighted the parlous state of LIAT’s finances which have resulted in the build-up of alarmingly large payable balances, including huge arrears to aircraft lessors and late payment of employee salaries. The stark reality is that LIAT cannot pay its bills and suppliers have been left holding the bag…..[snip]

There is no credible prospect of a turnaround at LIAT under the current board, management and policies. In fact, with regional travellers shunning LIAT because of repeated service failings and increased competition as eloquently detailed by Mr MacLellan, LIAT’s financial performance is likely to deteriorate further. Barbados will be forced to provide annual subsidies to LIAT in the form of transfers from the Treasury or loan guarantees at a rate that would pay the salaries of 500 public servants for a year.

Another casualty of Peak Oil? I submitted a comment suggesting just that and further suggesting that without a significant and sustained increase in world oil production, “There is no credible prospect of a turnaround at LIAT”, period.

It has dawned on me that we are now living through the slow motion train wreck that is, Peak Oil. If the masses only knew! I have been trying to do my part to raise awareness in my neck of the woods by using the words “Peak Oil” whenever I can or describing aspects of the problem or otherwise trying to arouse curiosity in the minds of readers of the comments section of my local newspaper. In a comment to a recent article outlining a suggestion by one of our opposition senators that the country “should consider limiting parents to two children”, I actually suggested that the person to whom I was responding do an Internet search for “albert bartlett arithmetic population and energy”, in order to better understand the point I was trying to make.

Surprisingly, it seems more of my comments are being passed by the moderators at that newspaper. Either they are softening their “see no evil, hear no evil, print no evil” stance or I am getting better at crafting my comments or a bit of both!

Nony

DC:

1. Nice work. I appreciate the effort to put together a framework. Doesn’t guarantee being right and we should consider other estimates also, but it’s good stuff. For one thing, it allows creating scenarios or even doing formal sensitivity analysis. (I saw a module in Excel that would do this for M&A valuations…you feed in the 10, 50, 90 estimates for different factors…and then the module actually calculates all the different combinations and gives you distributions and sensitivities.)

Nony

DC:

2. I think LaH gave you some very nice feedback on geology, but I would NOT feel bad about no matching his linearization. That’s based off of a couple years (somewhat arbitrary in decision) and is a very, very simple model. And we have seen the Hubbard linearizations be drastically off the last few years.

Nony

02/13/2014

I’m also happy to see you using some of the USGS research. I assume that you have seen this report on decline rates?

pubs.usgs.gov/of/2013/1109/OF13-1109.pdf

I have a little bit of a hard time with the super pessimistic estimates for ultimate recovery, given how off everyone has already been on the Bakken production rampup (just reading through old TOD posts and the like). Yeah, Continental is a promoter…but still there estimate is one to consider. And ND state was pretty brave (and early) in estimating the rampup that they did when everyone thought it would be crazy to get into the 700-1.2 MM bpd range. So, just be wary of the promoters…but also of people who hang out in a community called “peak oil”. 😉

Nony

http://www.cmegroup.com/trading/energy/crude-oil/light-sweet-crude.html

DC:

3. I think the price is OFF (much too high). You should use the futures price for oil in your model. After all, that’s the market guess…like a Vegas line. That’s how investors evaluate the play. Even if you think “the market is wrong”, you can still hedge. You can even look at the price guess thread in the peak oil forum and people are basically betting close to the futures estimate (not 140!).

I think price has a huge role in the Bakken and it is not impossible that we return to 40/bbl oil (not likely, but not meteor falling on your roof unlikely). IT IS PROBABLY MORE UNCERTAIN THAN THE GEOLOGY. We’ve even (in 2008) had one example of a price crash restraining drilling. You could also do a 10, 50, 90 for pricing (I think this can even be estimated with puts and calls), or just some scenarios. But really strongly urge you to think of price as one of the major drivers of the model.

I think WTI, Cushing is a good metric to use. Any decisions on export restraints or the dropping of them are already incorporated in that metric, so you don’t need separate factors. Also, the composition of the oil is almost identical to WTI. Obviously, there’s a transport offset (to Cushing), but that’s already in your model.

I’m really not a financier (just a civilian Internet commenter). But here is WTI futures chart:

It varies daily and there is some connection of current and futures price (e.g. current price has recently edged up a few dollars and so has futures price). But basically it has looked similar in shape for a few weeks. The “betting money” is that we have a gradual ramp down of about 20 dollars over the next 5 years. Note that these are nominal dollars, so since your model is in real dollars (I think), prices are actually declining even more than what is shown (in real dollars).

In any case, the 140 is really unrealistic. I like the Maguerri 65 (real) dollars much more. As far as extending later, I would just do a constant value after the last data point. If you want to get super fancy, you could try fitting some hyperbola or exponential and see if there is an asymptote.

D C

02/14/2014

Hi Nony,

Maugeri’s estimates are very optimistic, Mr Laherrere’s are very pessimistic, clearly you are more optimistic than I am. I think you may have not looked at the EIA’s AEO 2013 reference oil price scenario very carefully. In 2020 the scenario suggests $110/barrel (2013$) for Brent. Brent is the proper metric for the marginal barrel which is competing with imported oil at coastal refineries on the East and West Coast, the $140 per barrel is for 2033, I am pretty sure the futures market for 2033 is pretty thin and not likely to give a good estimate of future prices that far out (the EIA’s guess is likely to be wrong as well). Maugeri’s estimates for oil price and output are not credible in my opinion, prices of $65/ barrel in 2013$ in 2033, is that your call? So we will have to disagree, I actually think that if 2% worldwide growth or higher is maintained (on average) until 2033, that $140/barrel (2013$) is likely to be too low an estimate. In addition if prices were $65 per barrel (2013$) as Maugeri estimates then output will be much lower because wells will not be profitable, so keep in mind that the economics suggests that if the USGS estimates are correct (mean undiscovered TRR for ND Bakken/Three Forks of about 5.9 Gb) and prices are as low as Maugeri estimates, the economically recoverable reserves will be much lower than what I have estimated (assuming well costs do not fall below $7 million for drilling and fracking and transport costs, OPEX, and royalties and taxes remain close to present levels.

http://oilpeakclimate.blogspot.com/2013/09/update-to-north-dakota-bakken-three.html

The post linked above covers the TRR estimates which are for North Dakota only, which is about 79% of total US Bakken/Three Forks undiscovered TRR according to the April 2013 USGS Bakken/Three Forks assessment. See links in my blog post.

Also on futures markets see
https://www.stlouisfed.org/publications/cb/articles/?id=831

Futures markets do not do a very good job of predicting future prices. Consider betting lines for the superbowl in 2033, would anyone expect these to give them useful information? In the cornucopian world oil prices will always be lower in the future, in the doomer World the forecast is the same, but for very different reasons (economy in chaos, negative growth so very low demand for oil). Reality will be that oil prices will fluctuate up and down as they have since 2000, but the trend (5 year average real oil price) will be up, unless the more pessimistic scenarios (a permanent global economic crash where the economy is unable to find any substitutes for scarce fossil energy resources) is proven correct, my position is between these two fantasylands (cornucopian and doomer). Of course those with different positions than me believe that my position is a fantasy as well, (most here would call me hopelessly cornucopian).
Nony

02/14/2014

A. I’m well aware that a lower price means less oil extracted from the reservoir. Even though I have a cornie heart, I go with data.

B. Use the futures price as your baseline. It could be off high or low, but really…it is the betting estimate. We could be 40…or we could be 140, but why not make the central case based on the actual market indication?

C. If you think you are smarter than the market itself, bet on the futures themselves and just ignore exploration and the Bakken.

D. By making the base case, the futures market price what you do is constrain your model to looking at the play itself. Otherwise, you are confounding views on price (which depends on global supply and demand) into your model about a local play. Separate variables…it is better analytical method.

E. Go ahead and run a base, high, and low case. That way you can keep your “gut” feeling on price in the model, but segregated to a variance case.
Nony

02/14/2014

F. Also, re Magueri: $80 times 5 years of inflation is pretty close to 65 (although I’m not sure if Magueri uses real or nominal in his discussion). 140 is way out there and is in the opposite direction of the futures markets. Could it happen? Sure. But you are confounding a bet on global supply and demand (i.e. you think you are smarter than the markets) with the Bakken itself. Why not make base on the expected (i.e. futures) price. That is how anyone investing would make their decision.

Nony

http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/wacc.html.htm

4. The WACC you have there is very unreasonable. 15% is quite high. Sometimes companies will do things like requiring models to assume a high WACC (because they think other estimates will be padded). I was at a company that required using 12% (reality would have been 8% for them). But I definitely disagree on doing this for external estimates (and internal, but definitely external should reflect reality not an arbitrary WACC).

I’ve actually calculated this stuff from scratch (industry betas, term structure, etc.). But let’s just look at some already calculated industry WACCS:

Petroleum production is 8.6% (as of JAN 2013). And when you think about this play, the “years out” when the well is just declining and getting milked is pretty stable and risk free. And much more than exploration (but of course this play is getting pretty well understood, just by drilling…well at least the upper part).

Note that this WACC includes nominal dollars. I’ve seen top consulting firms make the mistake of using real dollars in the model estimates and then a normal WACC. Problem is that that includes implied inflation, which is implicit double discounting. Probably easiest thing is just to have the whole model in nominal dollars (then you don’t need to adjust the futures prices either).

Nony

5. Also, not sure if the USGS estimate includes Montana or not, but just something to watch for. Do you want to calculate all US Bakken/TF or just the ND part.

D C

02/15/2014

Just ND see my comment above and the sept 2013 post at my blog for more details.
Nony

02/15/2014

thanks. I figure it was in there, but I just didn’t read close enough. 🙂

Wanted to make sure you didn’t under/over count. 🙂
D C

02/16/2014

Nony,

Sorry for seeming defensive about your suggestions. Using the futures market as a guide over the short term makes sense, my analysis does this for Brent prices in real terms out to mid 2015 and then follows the AEO 2013 (roughly) out to 2030, I then extend the 2015 to 2030 rate of increase out to 2040. Actually my price scenario just takes the minimum AEO 2013 brent price (from 2019) and uses that price from Jan 2013 to Aug 2015, then follows the AEO scenario. The point being that futures prices are already being used over the short term, as far as I understand, futures prices for 2033 do not exist.

On the annual discount rate, your criticism was based on the WACC for all oil firms of 8.6 %. Two concerns would be that this would vary from small to large firms (generally being lower for smaller firms) and as you are WELL AWARE is not the appropriate metric anyway, we should be talking about MCC (marginal cost of capital). Generally the MCC>ACC, so we would expect the MCC for smaller oil firms to be higher than the WACC for all oil firms.
I agree with the criticism that using real costs and revenues eliminates inflation so the discount rate should be lower by 2.5 % for inflation (2.4% is the average inflation rate since 2002). Oil industry guy tell me 10 to 15% discount rates are usually used in the analyses they see, lets say 12.5% makes sense. Those analyses are no doubt in nominal terms so my analysis should use 10 % to account for using real revenues and costs.

The main point is that although I don’t agree with you in full, I appreciate the criticism, please keep it coming.
Nony

02/17/2014

I’ll get back to you with some more discussion on the topics. Not meant to be argumentative (or to drive you to change), but since I get the impression you want to chat it out.
D C

02/17/2014

Again, Thank you.

Not sure if you saw my last comment on page 3.

You have certainly given a lot of food for thought.

Your feel for prices beyond 2019 would be interesting. As far as I can tell there are no futures contracts beyond Dec2019. Would flat prices make sense to you from 2019 to 2040, continue the trend downward (if the real price was decreasing 2% per year just continue that out to the future? I would make this a low case because although you seem to place great faith in EMH and the implication that futures markets are the best guess, I have reservations. My understanding of the underlying analysis which leads to EMH is that future prices are modeled best by a random walk so that a guess that there will be no change in real price would make sense, that would be my base case and I could use the EIA’s AEO prices as a high case. You are no doubt aware that it is economists that do these AEO price forecasts.
Nony

02/17/2014

Yeah, I saw it. I’m actually trying to get caught up. I fired off without reading some of your work in full. I really don’t think it changes my remarks much, but figured if I really want to engage I should read up more. So I started with the USGS report. And then had to look up definitions on Wiki. And then realized I was looking at the fact sheet, not the report. 😉

For a lot of things, I can anticipate how I will respond: EMH and CAPM are really “orthodox”. Get a little discouraged with having to debate/explain them as (a) I’m not a finance/econ professor (b) you don’t have basic grounding, (c) somehow people think they know more than they really do…on the Internet. All that said, I probably should read the Fed paper before dismissing it (as a singular study versus the bulk of work, ignoring arbitrage arguments, in sample versus TRULY out of sample [i.e. after the paper written], and “do you wanna bet” remarks). 🙂

I am finding some interesting things that you will want to probably take a look at, regardless of freshman econ/finance discussion. Like the FEB Continental presentation has some good charts on the chemical makeup of the oil.

So…thanks for your kind demeanor, even if we have different views.

Enno

http://www.startribune.com/business/245552281.html

Great comments Nony, thanks for sharing your balanced view. I also fully support your idea on the importance of the oil price in all this. The USGS report was a very good read.

Nony

02/14/2014

Thanks man. You have good stuff too.

Nony

02/14/2014

December production was down. Not making any specific point on it (weather was bad). Just sharing the data point, hot off the Internetz.

Nony

02/15/2014

First greenfield refinery in US in 30 years.

http://www.vernwhittenphotography.net/wes1113/

Nony

02/15/2014

(pic 33)

Nony

02/27/2014

DC:

I may have to go back to work (not in oil), doing some contract consulting. Good for me/society, but much more stressful than playing on blogs.

Again, good effort and nice model. I meant to take the time to really read through it more and the supporting USGS stuff, but can’t commit to that. In particular, have not had time to think through/parse all the assumptions in structure of the model (what are the input, what are the outputs).

Some stray parting thoughts (do with as will, sorry if it is a critical “dump”):

1. Format:

a. Please hit the carriage return twice for new paras (to get a blank line). Also, separate the figures from text with blank line above and below. (I know a tiny nit, but really this stuff is kinda dense conceptually so walls of text make it worse).

b. Also, if you perhaps had some block diagram of the model, that would help with the understanding. (Maybe I’m tired, but I struggled to really read it…just fessing up!).

c. Also a tabular list of inputs: price (t), USGS case, costs, etc. And then a list of calculated intermediate outputs and the final output (I assume that is production (t)).

d. I think if you had a working Word document (updated as you tweak the model), that would be best. Hard to have all the info in a blog post. But you could link that (perhaps as a pdf, update as needed) and just write a teaser blog post. Or maybe a dedicated web page.

2. Structure: I think you need more of a full factorial of cases. There’s probably a way to structure this in Excel with a “control panel” (having the inputs) or with macros. I’m not a jock on that, but lots of people are. But 3 cases for price as well as 3 cases for geological would be good.

3. Price: I would do a high, low, medium case (at least).

a. A lower than futures based
b. Futures-based (For the extension, just leave either real or nominal $$ constant. Don’t care which)
c. Constant real, from now.
d. AOE medium (current)
e. An even more doomer one (perhaps AOE high?)

If 5 is too many, you could chose b, d, e.

Another option would be something based off of the EIA percent likelihood of price: note this is not an estimate, it’s actually derived from the market options themselves:

http://www.eia.gov/forecasts/steo/images/Fig1.png

And I don’t know how you extend it, but intuitively it should keep getting broader, but obviously less than linearly.

4. Real/nominal:

a. I think there is some confusion of real versus nominal with you. Options are valued in nominal, not real. They are bought on margin (very high leverage) and are settled with currency at the future time. It’s not like you buy the future option now with today’s dollars. And they’re valued the same way for both the short and long! So, the futures value for 2020 is the market’s guess at what oil will be priced in 2020 dollars.

b. Also, you had some comments about how real price changed over time in your model, but they didn’t seem to match your figure (which is annotated as “real price”).

5. EMH: It’s a very fundamental, philosophical thing. And sure, you can find all kinds of papers about people saying they found a flaw in EMH. But how interesting is it to publish a paper that says EMH is right? EMH is really orthodox and has some pretty powerful fundamental reasons behind it (basically arbitrage opportunities).

I would be leery of isolated industry studies saying that estimates (or constant real or whatever) outperform futures. I doubt they will extend into future performance, but are just happenstance from past statistical variation and cherrypicking/Bonferoni error.

I really doubt that even an EMH-naysayer econ professor would be willing to step up to the table and put $50,000 on FED paper assumption versus me taking the futures market. Market estimates are like democracy: screwed up, but better than the alternative.

6. 15% WACC is really high. I mean really, really high for almost any industry. (And it’s ~17-18% since you are using real revenue/cost in your model for the future years and thus double discounting inflation).

I don’t think it’s a good cost of capital for something as relatively defined as infilling the Bakken basin (which perhaps has an even lower risk than overall industry production…remember it is “mining” and is easily shut down or restarted if price moves…not like a deep water well.) If 8.6% is giving you heartaches, use 10% as a compromise. It’s sort of halfway between the oil production and oil services WACCs. And is within the “Rockman interval”. 😉

I already said that I don’t think much of company dictated discount rates. I’ve been at a big commodity company and they mandated 12% in their models…but it’s just some stupid big company rule. And NPV models have lots of levers for manipulation (watch me dream up synergies! ;)) You have to assume that over time, industries return their cost of capital no less/no more. That’s the way outside analysis should figure things. (I advocate it also for internal valuations: please see Copeland’s book Valuation which is the bible for this sort of thing and what it says about company discount rates.)

The Rock has seen some things go down in the industry, knows geology and oil companies. And is very kind to those less knowledgeable and I appreciate him on the boards. But he’s not a economist. So “Rock said so”…well it’s not his area of expertise.

15% is just really, really mean and harsh. Would not use it for a dotcom. Even if it is not changing the answer much, because of the high well decline rates, it just hurts me to see you using that as an assumption. Just shows a poor “feel” for this sort of thing.

I have never used MCC and it gets minor mention in Copeland’s Valuation or Brealey and Myers book Corporate Finance. I actually had to look it up (not a case of me knowing well)! Everyone uses WACC for valuations. For companies with optimized capital structures, MCC will equal WACC (I think, based on my recent reading…I am not a finance professor.) I would avoid using the term marginal cost of capital as it can lead to some logical fallacies (for example rating projects versus the source of financing…doing things like saying using own cash flow is riskless, etc.)

7. Reference price: I sorta like WTI since it also incorporates an estimate of “shut in crude”. And we really do have a situation of shut in light crude. There’s a differential and it is cheaper in the US, than worldwide. I think Brent spot prices are London (not sure).

Also, it’s just a bit closer in composition to Bakken oil. Although Brent is not bad, definitely a light sweet crude. LLS is the closest, chemically and in refinery yields. From what I read Bakken oil is actually slightly better in terms of value to the refiner than any of them, although LLS is close and we are talking a couple bucks difference.

There is a reference price at Minnesota pipeline entrance that is probably best proxy for the Bakken overall pricing. I think you have to pay some service for it, though. And I don’t know if it has futures. Really, the thing is complicated with some stuff going by train, pipeline, different coasts, etc. And it seems like companies try to move it around to whoever is best at the time. No strong preference here, just blathering.

8. Gut feel for supply and demand: I’m wishy washy, with no fire in the belly. More like humility of I don’t have a bias. (Not like the dotcoms or Enron, when I had very strong intuitions and insights that they were way overvalued, with some smoke and mirrors and MBAs thinking their poop did not stink going on.)

If you’d asked anyone in the 70s if we would crack OPEC and have 20 years of dirt cheap oil, they would have looked at you like you were a Martian. Remember Jimmy Carter’s synfuels project (kerogens)? The oil business has seen booms and busts before. (And from the Texas oil man’s perspective, a bust is not “world running out of oil”, it is “too damned much oil, price crashing”.) I’m not saying it will happen…and obviously a community that calls itself peak oil will be doubtful of that…but you have to factor some chance of it. We have seen things like this happen.

Then again, price is high now and has been for 6 years. That is very unusual/interesting/sucky for consumers.

The drop to 40 in 2008 is also curious. Was that a short term belief in long term supply being better? Or was that some aspect of OPEC slipping? Even with a global recession, one would not expect a 140 to 40 drop…because the current price needs to include the long term value of a depleting resource (IOW, better to leave it in the ground).

I really lack the analysis to say if OPEC is exerting market power (being a functioning cartel) or is pumping full out. IOW, could we have the repeat of 70s to 80s impact (especially with a supportive political regime…remember US oil production at the margin was a big factor in cracking OPEC and getting them to start cheating with each other.)

Then again, US tight oil came out of nowhere. Sure, Rock can say we knew there would be hard to get oil at high prices…but who drew a prediction curve showing 3 million Bpd (in about 3 years ramp) in the US? No one. So there are just uncertainties. Show me the Gail chart that had the LTO coming in. Or IEA or whatever. No one had it in mid 2000s…all they had was sands. And then Picollo predicted 150-225 M bpd, back in 2008ish. Rune predicted 600-700 M, back in JAN2013. So, just beware predictions from promoters OR from doomers.

Look at US gas. People were building import terminals for gosh sakes. And the long term futures were above 10. Now the long term futures are 4.50-5. Shale gas really has been a HUGE development and game changer. Much more so than LTO. (And that’s not oil…just my point is, things can change radically…there is uncertainty.)

I guess my gut is something like 75 real in the far off future. (basically extend the futures price). But with some risk of 30 or 150.

D C

02/28/2014

Hi Nony,

I appreciate you taking the time to do a detailed criticism/analysis. I don’t have time at present to respond in detail, but I may do so later so check back. A few things we continue to disagree about and/or I am not being very clear which I will mention briefly, but much of what you said about format and several price scenarios I agree with, there is definitely room for improvement.

On gut feel for future real prices we are on opposite ends of the spectrum. I would suggest that oil production will peak at some point (2020 to 2030 would be my guess), I also think that the decline may be relatively slow (1% or less for 10 to 20 years). If that is correct, I would be quite surprised to see real oil prices decrease once oil output stops rising. Perhaps you envision substitutes for liquid fuels making rapid progress so that they fall in price relative to petroleum powered transportation or that there is a radical adjustment in consumer preferences?

I guess the exercise boils down to whether one believes that the market is always right.

Have you heard the joke about the two economists walking down the street and they come upon a $100 bill on the sidewalk? As the first economists starts to bend down to pick it up, the second says, “Don’t bother, if it were real someone would have already picked it up.”

You seem to have sold faith that the futures market gives the best estimate of future oil prices, but during the Enron fiasco and dot.com bubble, you lost faith in the EMH, in the sense that you thought the market had it wrong. I also am unclear with how the EMH explains speculative bubbles, such as the housing market before the financial crisis. There are a lot of economic orthodoxies which seem very right, except in those cases where they are very wrong.

On real prices, I am not confused, the prices in my chart were for AEO 2013 (with the real oil prices adjusted to 2013$), I used the minimum price up to Dec 2015 and then picked a % increase which closely matched the reference scenario. I did this for simplicity because it did not look like price would be a limiting factor until around 2015, when Bakken oil might peak and lower supply was surmised to lead to higher prices (in line with EIA estimates).

I have more faith in the economists at the EIA than you.

Finally why use Brent rather than WTI? Much of the Bakken oil goes to the East and west coast by rail, for oil going to East coast refineries Brent prices are much closer to the refinery gate prices than WTI. So lets say Brent is $110 per barrel, and it costs $12/ barrel to ship the oil to the East coast refinery, the net revenue (before royalyies, taxes, and OPEX costs) is $98/ barrel (which is pretty close to the WTI price), to me the analysis makes more sense using Brent and transport costs, if I used WTI prices, I would eliminate the transport cost, it really is a wash.

I understand very well that futures are in current dollars. For example Brent future for Dec 2019 is currently about $91/barrel, that is about 6 years away and assuming 2.5% inflation over the next 6 years that would be $78/ barrel in Jan 2013$. If we extend this rate of decrease in price out to 2029 we get $45/barrel in Jan 2013$ and if we extend it out to 2039 we get $26/barrel (this is about a 5 % decrease in real oil price per year, which is the rate of decrease from Jan 2014 to Dec 2019.) Is this what you had in mind for the far future? Or does leaving the price unchanged from $78 in Dec 2019 to Dec 2039 make more sense? We get no guidance from futures markets beyond Dec 2019, and the market is pretty thin for Dec 2019, I think guidance beyond 2015 is pretty shaky.

See chart below for AEO 2013 price scenarios in May 2013$

D C

02/28/2014

Forgot Chart sorry

Nony

02/28/2014

looks like a good spread. Fine. Run the model and do the full factorial of price/geology. Should be interesting how date/amount of peak as well as total extraction vary for the 9 cases.
D C

03/01/2014

Hi Nony,

I will probably just do 2 cases, 8.4 Gb (USGS Mean) and 11 Gb (NDIC mean), I could do the 6 Gb case, but I think that is the least interesting. I did a quick scenario with flat real oil prices at $85/barrel and discount rate at 6.5 % (equivalent to 9% if nominal prices/costs were used assuming 2.5% inflation). ERR from 1953 to 2073 ends up at 6 Gb with those assumptions and my other assumptions (well costs fall to 7 million by 2017, royalties and taxes 26.5% of wellhead revenue{which is refinery gate price minus transport costs}, OPEX at $4/barrel and transport costs at $12/barrel.)

TRR in this case is 8.4 Gb and I used Enno’s data to develop a new well profile, peak is about 1.2 MMb/d in 2017 if EUR begins to decrease by June 2014 and reaches a maximum annual rate of EUR decrease of 14% in June 2015 (these are pure guesses, so far there is no evidence of EUR decrease, but we really won’t be able to see a decrease until 12 months after it begins.)

Nony

No worries man.

Nony

Here’s my gut (oops I mean Bayesian estimatation! ;)).

1. Basically just by eyeballing the 2012 versus 2013 increase. Even without the bad winter and a “normal DEC decline”, 2013 had lower absolute increase than 2012. So, something like:

2012 increase: 250k
2013 increase: 175k (if it had been only half the DEC decline)
2014: increase: 75k
2015 increase: -50k

So kissing 1 million bpd in some month of the late fall of 2014, but then finishing just below it. Top calander year, being something like 930 kbpd, 2014

EOG also notes the lower rise rate in Bakken (see their March investor presentation). Of course, they can preen since they basically kicked everyone’s asses by doing the Eagle Ford in 2010. But still, it’s just evident.

2. Some soft caveats in the other direction:

*Rig count for 2014 is starting to increase over 2013. This shows increased investment in the area and probably, given increase pad drilling is actually more activity than one would think. In JAN-FEB2014, I was seeing a pattern for rigs like 2014=2013=180, 2012=200. Last few days it seems more like 2014=190, 2013=180, 2012=200.

*The EOG presentation (and I sort of trust EOG to be a little less cheerleaderish than CLR/Hamm) talks about new techniques driving back some interest into the Bakken.

3. Based on gestalt of 1 and 2, I edge myself up and say monthly peak in mid 2015 of 1.1 MM bpd. Peak calendar year is 2015 at 1 MM bpd.

Nony

03/01/2014

Also, I do think Helms is telling the truth about what companies are planning in terms of continued drilling. Also, EOG presentation says 8 years of drilling remaining in their Bakken acreage.

Nony

I do think that the post-peak will be more plateu-ish than the peaker meme community would assume. Not a shark fin, not even a “other side of the bell curve”. Rationale:

1. Even with admittedly steep individual declines, the more time goes by, the more the population starts to include wells that are in lower declines. Imagine a single year of drilling 1500 wells and then stopping cold, versus 10 years of drilling 1500 wells and stopping cold.

2. There will be some tapering in of less drilling and of drilling marginal wells (lower IPs). Not a stopping cold.

3. Refracking of select wells. I’m not sure what the criteria will be (the big wells justifying it, or doing them checkerboard to stimulate several wells or some feature of local geology). But I think it will be something like some of the wells justifying refracks…not all, not none.

Nony

In terms of price assumptions, I think the play is more take-off limited now or just logistics limited. So, while there is price sensitivity, my intuition is that it may not be as readily observed as people think. If oil goes to 150, they can’t ramp up to 50% more production. Similarly if we drop down to 80, a lot of the infill drilling is still profitable, just not as high return. At 60, you would see very significant contraction of drilling. At a prolonged 40, the whole play turns off (we sort of saw this in 2008 already).

Nony

03/01/2014

take-off, logistics, and geology limited.
D C

03/01/2014

Hi Nony,

My scenarios do account for the lower decline rate of the older wells. In fact the well profile that I use, extends out to 100 years and assumes that once a well starts producing it never shuts down (so it is a little on the optimistic side in that sense). The TRR models assume that economics is not a factor, but that eventually the drilling stops at about 45000 wells or so, this could be increased, but would have little effect because I am assuming the USGS estimates are correct, more wells would just result in more of a new well EUR decrease to keep total TRR at the USGS mean estimate.

I don’t assume higher prices result in a take off in production, just that the number of wells added would decrease more slowly if prices were higher because drilling would be profitable for a longer period. Once profitability per well gets below about $1 million (2013$) per well the number of new wells added are decreased gradually so that profits remain positive for each average new well added. The only way a plateau could be maintained is for prices to move higher, you think that prices will not move higher (judging by futures prices), so a plateau would be unlikely unless the refracking of wells becomes important. This is certainly a possibility, I believe it is probably on the order of 3 to 5 million dollars per well and the question is how much of a return in increased output would result, I do not know the answer to this question, perhaps some of the investor presentations you have looked at would shed some light. My models do not account for this effect and I do not know if it will be of great significance going forward. I might be able to get Enno to dig some of this (refracking data) out of the NDIC data, so it is definitely worth looking at.

On the infill drilling being profitable at $80/barrel, that depends on what the EUR is for the infill drilled wells, at $85 (2013$)/barrel you need about 200 kb 30 year EUR for breakeven at a 6.5% discount rate, this might work in the best sweet spots, but there is not unlimited space for infill drilling. At $60 (2013$) I think things shut down.

Note that the parts of my curves that are sharkfin-like are where profits are zero for new wells. I assume (perhaps incorrectly) that these companies are not in business to lose money, if expected profits ore zero or negative, new wells will not be drilled.

I could try to create a case where a plateau is reached, but it would require prices that are not at all in line with the futures market. In fact the AEO reference scenario that you thought was unreasonably high, did not result in a plateau, though perhaps with a lower discount rate it might.

Nony