The EIA STEO was published recently the estimate below is based on data from that report and statistics from the EIA International Energy Statistics. The EIA expects the 2018 peak for annual average World C+C output will be surpassed in 2025.
Beyond 2024 the EIA forecast looks optimistic to me, my expectation is slightly higher output in 2024 (annual average World C+C output) of about 82100 kb/d (about 250 kb/d higher than the EIA forecast) and about 82400 kb/d in 2025 (roughly 1250 kb/d less than the EIA forecast).
There have been a number of revisions since last month which are detailed below.
Oil, diesel and gasoline prices are lower and natural gas prices higher than forecast last month.
It is not clear why these relatively flat Brent prices would result in the strong increase in production in World C+C output from May 2024 to December 2025 that has been forecast by the EIA, in my view flat output would be more in line with this oil price forecast.
The EIA expects slightly less output from OPEC+ than the targets announced by OPEC+ (shown in gray in the chart above).
The chart above contrasts the strong increases in output from non-OPEC+ producers relative to OPEC+ especially since 2022 and the srong increase is forecast to continue in 2024 and 2025.
Rising natural gas prices might help profitability in the Permian basin, but note that typically Waha Hub prices which reflect the natural gas price that Permian producers may see is often $1 or more lower than the Henry Hub price. Note also the huge fluctuation in natural gas prices from almost $9/ MMbtu in 2022 to under $2/ MMbtu in 2024, 4.5 times lower. For oil at $40/b a similar level of change would mean an increase to $180/b, natural gas prices are much more volatile than oil prices.
Stock draws are expected in 2024 followed by stock builds in 2025 based on the EIA forecast in June.
OECD commercial liquids inventory is expected to be close to the 5 year low over the forecast period.
Note the difference in symmetry between the 95% confidence intervals in oil price vs natural gas price. For oil the range for December 2025 is 40 to 120 per barrel with the STEO best guess at about 80/b. For natural gas the range is 2/MMbtu to 9/MMbtu with the STEO best guess at about $3.50/MMbtu. There is much higher upside variability for natural gas compared to oil (the equivalent for the high estimate for oil would be about $206/b).
In the US, energy expenditures as a share of GDP have been generally decreasing over the 2005 to 2024 period.
This chart indicates the dominance of the Permian basin especially since 2021, with most of the tight oil increase from 2021 to 2024 coming from the Permian.
The Permian also played an important role in keeping overall shale gas output pretty flat since the end of 2022 (with other basins as a group showing decreasing shale gas output since the end of 2022.)
The chart above includes NGL output with dry shale gas output and also includes conventional natural gas from each region.
The chart above shows that for the US C+C output since 2021, most of the increase has been from the Permian basin, it also shows that future annual rate of increase from April 2024 to December 2025 is expected to be smaller than the Jan 2021 to April 2024 period. I believe the EIA forecast for the Permian Basin may be a bit optimistic. The chart above includes both tight oil and conventional C+C from each region.
The scenario above assumes the EIA forecast for oil and natural gas prices from April 2024 to December 2025 is correct, it is for Permian tight oil output only rather than regional Permian C+C output (C+C output would be roughly 400 kb/d higher than tight oil output from 2023 to 2025 in the Permian basin.) Compared to the EIA forecast for the Permian basin, this scenario’s centered 12 month average is about 247 kb/d lower in 2024 and 632 kb/d lower in 2025 than the EIA’s forecast for Permian tight oil output in 2024 and 2025.
Great work Dennis!
https://www.youtube.com/watch?v=BRqIezbVMHM
7 minutes
The Energy Crisis in Europe due to Russia
juicy info in here!!
Ron – can the above comment be deleted and the user banned please.
+1
Banned for what….posting a link to energy content on an energy blog?
Sorry Andre this comment refers to “R” who had posted an anti-Semitic comment in reply, not to you. Their comment has since been removed. Dennis or Ovi, can you clean up this message thread so it doesn’t look like we are referring to Andre? Thank you!
no worries mate!
What? No cliff?
Kidding aside, on the US ‘energy expenditures’ do you know what that includes?
Just purchases of fossil fuels?, or would it include uranium, photovoltaics, transformers, LNG export terminals, sand for fracking, and the other myriad expenses that comprise energy sector costs?
That chart indicates a dramatic drop in national energy expense as a percent of GDP.
Maybe other component expenses of national GDP have skyrocketed, such as obesity treatment and political advertising.
Hickory,
Pretty sure it is all spending on energy as a percentage of GDP.
Hickory,
I checked on this and yes it it expenditures on all energy (not just fossil fuel). Here is a chart with data from 1970 to 2022 (data for 2023 not out yet) and no data before 1970. The straight line is the OLS trend which probably is not meaningful. The chart is for the US only.
Thank you Dennis. This will surprise many people.
A take home message is that ‘GDP’ includes many things, some components having grown very fast.
Hickory,
GDP includes all goods and services produced.
See https://fred.stlouisfed.org/release/tables?rid=53&eid=14897#snid=14923
for some of the detail.
If you are concerned about financial and insurance and recreation expenditures, the chart below subtracts those from GDP and looks at energy expenditures percentage as a share of this “alternate GDP”
I am assuming this is in part due to efficiency. We have managed to get more unit output per unit Joule.
To get an exact understanding, i guess one has to look at how GDP is measured, which may or may not have changed over time, which may also give us the wrong impression and worldview.
Iron Mike,
The ratio is nominal energy expenditures to nominal GDP so inflation measures are not part of the equation. Statistics are never perfect, this is just the best the US government has to offer, data goes back to 1929 for US GDP, we only have energy expenditure data from 1970 to 2022.
For GDP see
https://apps.bea.gov/iTable/?reqid=19&step=2&isuri=1&categories=survey&_gl=1*14nkq3v*_ga*MjY1NDE0MTAwLjE3MTQwNzE2MjE.*_ga_J4698JNNFT*MTcxOTU4ODM0Ny45LjEuMTcxOTU4ODY3Ni4xOS4wLjA.#eyJhcHBpZCI6MTksInN0ZXBzIjpbMSwyLDMsM10sImRhdGEiOltbImNhdGVnb3JpZXMiLCJTdXJ2ZXkiXSxbIk5JUEFfVGFibGVfTGlzdCIsIjUiXSxbIkZpcnN0X1llYXIiLCIyMDIyIl0sWyJMYXN0X1llYXIiLCIyMDI0Il0sWyJTY2FsZSIsIi05Il0sWyJTZXJpZXMiLCJBIl0sWyJTZWxlY3RfYWxsX3llYXJzIiwiMSJdXX0=
for energy expenditures see
https://www.eia.gov/totalenergy/data/browser/?tbl=T01.07#/?f=M
Thanks for the link Dennis, very interesting datasets.
Just a sidenote according to vox.com:
The basic formula — consumption plus investment plus government spending plus net exports — never has, but the US in 2013 made a big change to its GDP accounting when it added both artistic creation and research and development to the total, which it had never done before. The change came because the international standards, the System for National Accounting, changed in 2008 and the US wanted to make sure its statistics would remain comparable to the rest of the world’s (this means the rest of the world has also or is also making these changes).
Those changes started including intangible things like the production costs on movies or a drug company creating a new treatment as investment, rather than intermediate costs. In addition, the BEA also started counting defined benefit pensions differently, adding them to GDP when workers recieved the payouts, not when employers paid money in.
Not sure if removing those components would have much of an impact but good to know eitherway.
Dennis,
Regarding your second link EIA’s primary energy consumption, I’d like to overlay that data with world birthrate, i think there might be some correlation there.
I’d go along with efficiency gains, but also include the exporting of so much heavy industry and manufacturing as well. It’s the latter that means we should be looking at the world as a whole for this type of number.
On the efficiency gains, they give us a false sense of everything being OK providing we keep increasing efficiency. It’s a trap..
Let’s assume we get so good, ‘we’ get to 100% efficiency, or whatever physics theoretical maximum efficiency is for something simple like crushing and grinding rocks to liberate the minerals, at some point in xx years time.
10 years later we will be mining lower grade ore, meaning more rock to crush and grind to gain the same quantity of minerals, so more energy has to be used to do it, an ever increasing quantity of energy to produce the same quantity of minerals as the years go by. In the long run it’s physically not possible to continue like this on a finite planet.
Hideaway,
You are generally correct and you are looking at it from a physicist/scientist viewpoint. Lets expand on that perspective and add an economist/technologist viewpoint:
You mention 10 years later we will be mining lower grade ore. This is true, but human ingenuity will allow the lower grade ores to be mined efficiently. So technology will definitely play a factor and kick the proverbial can further down the road.
These topics are complex and multifaceted and it doesn’t help to look at it myopically. Yes, the fact is there is a limit, but that limit maybe significantly delayed due to human ingenuity.
Hideaway,
I agree looking at it at the World level is better as high energy use industries have mostly left the US.
The best we can do at the World scale is energy consumed per unit of real GDP because we don’t have the energy expenditure data for the World, only energy consumption in EJ or TWh. Chart below gives World Fossil fuel consumption in MJ per 2015$ of real GDP from 1965 to 2023 using World Bank data for real GDP
https://data.worldbank.org/indicator/NY.GDP.MKTP.KD
and Statistical Review of World Energy data for Fossil fuel consumption
https://www.energyinst.org/statistical-review
I agree that for energy in general there will be a physical efficiency limit, but for fossil fuel, energy use might eventually be reduced to zero (some fossil fuel may continue to be used for material inputs for some products).
I also agree growth cannot continue indefinitely, but I expect the demographic transition will result in a peak and decline in population and that eventually we might reach a steady state (real GDP growth of zero.) Though that would be 100 to 200 years in the future.
Iron Mike,
I agree likely due to increased efficiency (if underlying data is roughly correct). Chart below looks at energy consumed (in thousands of btu) per dollar of real GDP (2017$), in this case an inflation index is used (chained 2017 $) to convert from nominal to “real” GDP. Data from 1949 to 2022. Note that CO2 emissions per dollar of real GDP (2017$) (a rough proxy for fossil fuel consumption) decreased by about a factor of 5 from 1949 to 2022 where energy consumption rather than the factor of 3.27 decrease in Energy consumption per $ of real GDP (2017$).
Thanks Dennis,
If you have the data, can you overlay U.S population growth rate with that dataset to see if there is a correlation. I don’t have the dataset to do it myself.
Iron Mike,
See https://www.eia.gov/totalenergy/data/browser/?tbl=T01.07#/?f=M
for the data, it is unclear what you are looking for.
Dennis,
Have a look at the chart in the link, scroll down. Annual growth rate of the population for U.S.
https://www.macrotrends.net/global-metrics/countries/USA/united-states/population-growth-rate
There seems to be a correlation with US energy consumption. Seems like a possible causation too in my opinion. As the birthrate drops so does the energy consumption. Could be another factor on top of efficiency. I could be wrong though just going off shape of the curves.
Ill try to get the datasets and see the R^2.
R^2 is ~ 53%
I think it could still be a possible causation as to less energy per unit GDP, declining birthrates. I could very well be wrong though.
Iron Mike,
An aging population could also reduce energy consumption perhaps. US population vs energy consumption, correlation from 1982 to 2007 is good, but 2008 to 2022 not very good, R squared about 0.083.
We do have a significant immigration component to population in the US…not just domestic birthrate.
Dennis,
That’s a great point and that demographic change would be another factor.
Hickory,
My bad I shouldn’t have wrote birthrate, it is actually the U.S population growth rate which includes immigration.
As far as energy embodied in sand for fracking, probably only included as far as costs passed on when energy products is sold, the expenditures are on final energy products such as natural gas for heating a home gasoline or diesel at the pump or electricity sold to final consumers.
https://www.oilystuff.com/forumstuff/forum-stuff/down-to-one-bench-in-saudi-america
Saudi America down to one bench
Very interesting article. Did Exxon got played? Early signs are “yes”.
I am not sure if the minor benches can sustain the plateau, but I am sure they will be drilled, even if production flows are disappointing. As the wells decay, the more wells you have per pad, the more late-stage production per pad you have, the better utilization of the pad infrastructure: gathering, separation, transport, water disposal, etc.
Some companies are drilling up to 30 wells per pad in some areas, the economics is favorable…
The Short-Term Energy Outlook for the next two or three months is based on estimated production. After that, it is based on estimated demand. The EIA just assumes whatever the demand is, production will rise to meet it. Therefore, the predicted production is based entirely on the predicted demand.
Perhaps they are right. But I don’t think that is the case.
I have worked in an Energy Organisation.
We used linear regression to forecast electricity demand and capital/operational budgeting.
linear regression doesn’t work well if things aren’t linear!
None of us were electrical engineers or boots on the ground.
The higher ups had the ability to override the forecasts.
And u can’t change the forecasting algorithm because u think u have a better idea!!!!
Try suggesting a non-normal generalised polynomial regression and hear the gasps throughout the organisation.
And don’t forget all the stakeholders using that data to form decisions. Companies are doing their budgeting on what the EIA says!!!
The EIA may be inaccurate, but they are not sinister. Atleast not the guys/gals actually doing the work and not playing politics
I’ll bet atleast 95% of people working at the EIA don’t know what Peak Oil is or it is a bullshit conspiracy. And those that do … are afraid to speak up.
Ron,
I agree.
Thanks Dennis,
As usual i find the EIA’s forecast as odd. Looking at the second chart, the table shows that in 2025 US GDP to drop down to 1.6%. While brent crude and gas prices are higher on average than the previous year.
So they are seeing an economic slowdown but higher energy prices. I find that rather peculiar.
Also as you noted, they see a much higher volatility in gas prices compared to oil. Which again for me is a bit strange. I wonder if they have any possible causation as to why this is in their modelling.
Iron Mike,
You are welcome, thank you for the feedback.
The price for Brent Crude would be determined by the change in World GDP as it is determined by the World oil market, for natural gas they might be expecting that there will be higher exports of LNG, the 95% confidence interval is based on the futures market data for Brent and Henry Hub markets as far as I understand.
Generally at the EIA they think in linear terms so if last year plugin vehicle sales increased by x million units (I don’t have that number) they would tend to assume that same increase would occur in 2024 and 2025. I tend to see past exponential growth as continuing so reach a different conclusion, but I may well be wrong. Generally plugin vehicle sales have increased more than I have estimated in the past, perhaps in this case I am over estimating.
In June 2019 I did a “realistic” EV scenario which had the World plugin light duty vehicle fleet at 37.5 million at the end of 2023, current estimate is about 42 million for World light duty plugin vehicle fleet. I was not accounting for the pandemic in my model as I had no idea of the impending pandemic back in the summer of 2019.
914 out, with a moderate monthly increase, for APR24 data.
https://www.eia.gov/petroleum/production/
OH back in front of LA. (MT and KS not that far behind, also.)
I still would put (quite literally) money on ND breaking 1.3 before year end. You all underestimate its recovery from winter, often.
“Rigs targeting oil in the Permian Basin fell by 3 to 303, and rigs in the Eagle Ford also lost 3 to 44.”
https://seekingalpha.com/news/4120700-us-drilling-rigs-drop-for-eighth-time-in-10-weeks-in-baker-hughes-count?source=content_type%3Areact%7Cfirst_level_url%3Amarket-news%7Csection_asset%3Amain
The Rig report for the week ending June 28
– US Hz oil rigs decreased by 7 to 431. This is the eight consecutive weekly drop, down 28 from 8 weeks ago and the lowest since January 22 when the rig count was 434 and rising after the Xmas break.
– The Texas rig count dropped by 5 to 236 rigs and is down 90 from January 2023 . Of the 5 rigs, 3 were from the Eagle Ford basin and there were 2 more from the Texas Permian.
– Texas Permian dropped 2 to 190 while the New Mexico Permian dropped 1 to 102. The Texas Permian is down 54 rigs from the May 2023 high of 244.
– In Texas, both Martin and Midland counties were unchanged at 30 and 18 respectively.
– Eagle Ford dropped 3 to 43.
– In New Mexico Lea was unchanged at 50 while Eddy dropped by 1 to 51.
– NG Hz rigs were unchanged at 86. (Not shown).
Frac Spread report for the week ending June 28.
The frac spread count dropped by 9 to 237 and is down 35 from one year ago. It is also down by 35 spreads since March 8. The overall trend in frac spreads appears to be down.
This is a big weekly drop and raises the question; “What is behind this big drop in such a short period?”
US April Oil Production
Production rose by 72 kb/d to 13,248 kb/d with Texas being the biggest contributor, 53 kb/d. On a YoY basis, April production is 598 kb/d higher.
It is difficult to understand how Texas is increasing production while dropping rigs and fracs.
Ovi,
We don’t know the frac count for Texas, Texas rigs have been pretty flat from Aug 2023 to May 2024, if there is a 7 month lag between drilling and first flow the drop in rig count starting in May wouldn’t show up until November. Surprising we have seen a drop in New Mexico output this month.
Dennis
Yes I realize this won’t show up till the end of the year. Rigs are real time and this makes me wonder if the drop in rigs is due to a one time event in Texas, WTI price or fewer Tier 1 wells????
As for NM, I wasn’t surprised. Eddy’s drop overwhelmed Lea’s small increase in April. See attached Eddy chart which will be part of the next US update. The two month methodology is working fine for NM but not so much for Tx. More on that in the upcoming US update.
“Mergers and acquisitions in the U.S. shale oil and gas space could lead to a decline in production, industry executives said in the latest edition of the Dallas Fed Energy Survey.”
https://oilprice.com/Energy/Crude-Oil/Shale-Consolidation-Could-Reverse-Output-Growth-Trend.html
The current production is coming from oil wells fractured long ago when the frac spread was increasing?
Jean
In the above chart, the rigs data has been shifted forward by 8 months since I think that production should follow rigs provided only a few DUCs or preferably none are being used. That is a good assumption for the last year. We should expect a further drop in production for the next few months.
Production should then increase starting around July. The interesting question is “Will production get back to the February 811 kb/d level. I don’t think so because of the increasing legacy decline.
“It is difficult to understand how Texas is increasing production while dropping rigs and fracs.”
I often see this misunderstanding, which is annoying given the number of years you people (trademark Ross Perot) have been amateur analyzing this stuff.
1. AT a given level (say 5.5 MM bopd), there is a certain base decline and to offset it a certain number of new completions needed. (Yes, the base decline itself may vary over time and the new production per completion may change also. However, in near term, it is static…what matters is the number of new completions. The DPR sort of looks at things this way. Although they use rigs (not frac suites or completions), they are inherently looking at new oil versus base decline. The important realization then is that if we are ABOVE that level X for base replacement (i.e. in a growth regime), then a drop in rigs (or frac suites or completions) will not lead to a drop in production UNTIL the amount of new oil/month drops below base decline.
It’s as if I need the gas pedal at halfway, to maintain 60mph on the highway. If I floor it, I will start accelerating. If I then (rapidly and before major changes in my speed itself), move the pedal to 3/4 floored, I continue to speed up! This is because we are still in the growth regime. Similarly, if I had the pedal not depressed at all…at 60 and then pushed it down halfway, I would NOT stop decelerating. Since I’m still giving less gas than the equilibrium (half in this example) setpoint.
This is very clear if you have ever worked with control circuits or physics or differential equations. It’s something Enno would grasp intuitively. But I see hoi polloi confound themselves with it often (and from both sides of the aisle).
2. You’re looking at APR oil and recent frack count. although I grant you, it’s been trending down for a while. But still you can’t compare this weeks frack count with two month ago oil.
3. There is some slop, some play in the system. A jitter in fracs (or rigs or completions) won’t immediately show up in new oil. It’s like wiggling the rudder on an ocean liner, you need to wait for the response. It’s not a sports car.
4. There are still efficiencies, improvements growing in the system. Even if well productivity is not going up (although it went up for several years, like 2014-2019 when peakers said we were already used up all the good rock), the speed of completions and drilling has still been increasing. So a given spread does more (if it works faster).
ANONYMOUS
I will have charts in the next US update that will clarify why I raised the question regarding Texas production.
Make sure you give Ohio it’s street cred. 😉 😉 And what is going on with MT? Interesting.
Dennis,
I have a cool chart I want to post, but the site is just rejecting the post. Is there a known bug? Fixes?
I believe it has to be under 60kb in size. You can change the file size by using the resize button in microsoft paint.
Kdimitrov,
You have to reduce the size of the chart to less than 55 kB, sometimes saving file as gif or jpg will reduce size. I typically copy to a spreadsheet then adjust pixels to reduce size, usually jpg format works best.
Looking at the 2024 Statistical Review.
https://www.energyinst.org/__data/assets/pdf_file/0006/1542714/EI_Stats_Review_2024.pdf
Bearing in mind that oil consumption grew and total liquids consumption grew by over 2 million barrels per day. Coal consumption grew by over 200 million tonnes and gas consumption grew.
Renewable energy grew by 5 billion Gw hours yet failed to reduce any of the above.
How many additional wind turbines would need to be built each year to start to reduce the amount of fossil fuels.
Let’s assume that hydro continues increasing at current rate for a few years until governments realise that the land flooded is too precious. We can assume that solar increases at the rates it has been.
If oil production peaks and starts falling by 1mb/d some people have calculated that 15 thousand 4Mw wind turbines would be required just to make up that shortfall.
As for electricity consumption
https://ourworldindata.org/grapher/per-capita-electricity-generation?tab=chart&country=USA~IND~CHN~AUS~OWID_WRL
Obviously electricity consumption would have to increase substantially in order to reduce coal and gas consumption also.
Nonfossil energy, whatever the source, will offset fossil fuel combustion only if civilization decides to use less energy. Or if fossil fuel become too expensive in relation to other energy.
And of course combustion will decline simply due to depletion…coming to your neighborhood soon.
Peak Global Combustion in just about 10 years.
Hickory
30 years too late
https://www.unep.org/resources/emissions-gap-report-2023
Tim,
Focus on crude plus condensate for liquids consumption, a good portion of NGL goes to the petrochemical industry to make plastic and other polymers (polyester etc) which are not necessarily a part of energy consumption, but become material products. There may be some energy emissions in these processes, but some of the carbon is bound up in the material output. the major proportion of energy from liquids fuel is in the crude plus condensate portion which gets refines into light and middle distillates and bunker/residual fuel. This accounts for about 88% of crude plus condensate with the rest as byproducts used in the refinery (petroleum coke and still gas) or other byproducts such as roal oil, asphalt, wax, lubricants and other products.
Chart for World consumption of light middle distillates plus fuel oil (aka bunker fuel).
From 2018 to 2023 the average annual increase was about 51 kb/d for the World. See Oil-Regional Consumption tab of Stat Rev of World Energy spreadsheet.
https://www.energyinst.org/__data/assets/excel_doc/0020/1540550/EI-Stats-Review-All-Data.xlsx
Over time as more BEVs and plugin hybrids replace ICEVs liquids use will peak and then decline.
https://arxiv.org/pdf/2401.16447