Peak Oil Demand Is Coming, But >4 Mmb/d Of New Oil Supply Will Be Needed Every Year To Replace Declines To Get There
Nov 6, 2017
We buy into the narrative of peak oil demand, believe it is inevitable, its visible and will happen before 2030. Peak oil demand will be from the cumulative impact of a number of factors including EVs, battery/storage, LNG for power, LNG for transportation, increased energy efficiency, etc. But the peak oil demand narrative forgets the most basic fundamentals of oil – industry has to add new oil supply every year to replace declines just to keep production flat. Even after today’s big oil rally, long dated strips are still under $52 from 2020 thru 2025. We don’t believe long dated 2020 thru 2025 strips are predictive of future prices or indicative of the marginal supply costs to add 4 to 5 million b/d every year in 2020 to 2025 or to add >3 million b/d every year once peak oil demand is reached and is in plateau. We believe these marginal supply costs are significantly higher and >$60. We believe oil can quickly move to a base of >$60 with this supply challenge and there will be longevity to this call as markets appreciate this challenge and that the marginal supply cost to add this much new oil production every year is well over $60. Peak oil demand won’t take away from the challenge to add significant new oil production every year.
Let’s move beyond the debate IF peak oil demand will happen – it’s going to happen. Last week, we had the opportunity to discuss long term oil issues at a true strategy session for a producer. The first question we had was if we had drank the kool-aid on EVs and how EVs would crash oil demand. We replied more broadly, saying yes, we have drank the kool-aid on peak oil demand, believe it is clearly visible and driven mostly by technology. Right now, the narrative is all about EVs crashing oil demand, but our thesis is that peak oil demand doesn’t happen from one event but from the cumulative impact of a number of mostly technology driven items that are leading to substitution of oil by electricity and natural gas. It isn’t just EVs, but includes items like batteries/storage, LNG for transportation, LNG for power, etc. We believe peak oil demand is visible and coming, and before 2030. But EVs are the most visible and topical, so wanted to include some added comments on EVs.
The formula for calculating EV’s impact on oil demand is simple, it’s the assumptions that drive a wide variance in estimates. There is a common starting point for any estimate of future EVs – its about 2 million EVs, with EV sales in 2011-2016 being EU 33%, US 33%, China 26% and Japan 8%. The formula to estimate EVs impact on oil demand is a relatively simple formula with 3 key factors – how many internal combustion engine (ICE) vehicles are replaced by EVs, in what regions of the world are EVs replacing ICE vehicles, and what is miles per gallon or litre of the ICE vehicles being replaced. The assumptions on each of these three components leads to a wide variance in estimates. For example, EVs sales by region is important as the average mileage driven in Europe or Japan is substantially less than in the US/Canada. Replacing an ICE vehicle in Japan has a lesser impact than replacing an ICE vehicle in the US. How will an ageing population impact average miles per vehicle driven each year? The US Federal Highway Administration statistics show that ages 20-34 and 35-45 drive, on average, approx. 15,000 miles per year, but that declines to approx. 12,000 miles per year for ages 55-64, and approx. 7,600 miles per year for over 65.
No surprise, there is a wide range of EV estimates. Fortunately, there are many groups who put time and effort into looking long term at what EVs and their potential impact on oil. These groups provide good color on the key items for their forecasts. On the optimistic side, Bloomberg New Energy Finance (July 6, 2017) “predicts electric vehicles (EV) will displace 8 million barrels of oil a day by 2040, or about 8% of current global crude demand”. On the pessimistic side, ExxonMobil said in its Oct 27 Q3 earnings call “Now our energy outlook assumes on light-duty vehicles that we will reach a peak in gasoline demand and come down, and that’s driven by two things. It’s driven by efficiency of ICE, of internal combustion engines, as well as greater hybrid and EV penetration into the marketplace. So the bottom result of all this is, is that we do look at variations in the peak or in the EV penetration. And if you look at our forecast right now, we have by 2040 the fleet is about 6% EV. If you were to increase that by 50%, that would have a liquids demand impact of about 0.5 million barrels per day. So not substantial when you think about overall oil demand of over 100 million barrels per day at that point, but certainly, something that we keep a very close look at as we think about our research and development activities, as well as building the strategy.” That’s a huge difference between Exxon and Bloomberg for the EV impact in 2040. BP Amoco uses a 2035 data and estimates there will be 100 million EVs (~6% of total vehicles) in 2035 and that will reduce oil demand by 1.2 million b/d in 2035. Finally, the IEA uses a 2030 data and its June 2017 “Global EV Outlook 2017 Two million and counting” wrote “Assessments of country targets, OEM announcements and scenarios on electric car deployment seem to confirm these positive signals; indicating that the electric car stock may range between 9 million and 20 million by 2020 and between 40 million and 70 million by 2025”, and included a reference case for 2030 between 55 and 60 million EVs in 2030. .
EVs will continue on an accelerated sales curve, the slope will be determined by infrastructure” We have also drank the kool-aid that EVs will continue to have an accelerated sales curve. But we think the infrastructure isn’t there over the next several years and the biggest acceleration doesn’t come until the late 2020’s. We believe the key to the big acceleration of EVs is the grid and the ability of the grid to deliver power to support owners looking at EVs like ICE vehicles and not have to plan an overnight “fill up” in advance. Currently, >95% of EVs are charged at home overnight. The grid will have to be prepared to supply high quick demand for power and do so during peak daytime hours to support “fill up” convenience similar to ICE vehicles. And it has to do so in a world of increasing clean energy in the energy mix that isn’t necessarily leading to affordable and reliable electricity. Ontario has had a well publicized deferral of some clean energy initiatives in 2016 (some say driven by the upcoming 2018 elections) to try to control electricity prices. On Oct 17, Australia announced a “National Energy Guarantee to deliver affordable, reliable electricity” that also deferred some clean energy pushes also to control electricity prices. The other big timing wildcard for acceleration, at least in the US, is if Trump and the Republicans ultimately end up with the early termination of the $7,500/unit maximum federal tax credit for EVs.
50 million EVs in 2030 is probably a reasonable case. We tried to come up with a reasonable expectation based on interpreting the logic for the key assumptions. We might have moved to the IEA or slightly higher ie. 60 to 70 million EVs if it not for the Australia/Ontario on the challenge to deliver affordable and reliable electricity, our concerns that it becomes more challenging to accommodate “fill up” like ICE convenience during peak grid hours, and the potential slow down in growth rates in the US if tax credits end early. A 50 million EVs in 2030 is not inconsistent with 100 million in 2035 as our assumption is the acceleration in EVs happens in the late 2020’s.
At 50 million EVs in 2030, we see EVs reducing oil demand by 0.5 million b/d, but it would be 1.0 million b/d at 100 million EVs. This compares to expected global oil demand between 105 to 110 million b/d. Our key assumptions for the impact on oil demand are as follows. We are like others and assume continued efficiency gains in miles per gallon (we use 15% better in 2030) and less miles per vehicle driven per year with urbanization and an ageing population (we use 25% less miles per vehicle in 2030). We have assumed that US EV growth slows down relative to other regions on the potential of an early termination of the EV tax credit, which means the US% of global EVs in 2030 is only 20% versus 33% of EV sales in 2011 to 2016 sales.
Estimated Oil Demand Displaced By Electric Vehicles in 2030
Source: Stream Asset Financial
The peak oil demand narrative overlooks the need to add >3 mmb/d of new supply every year to replace declines even after peak oil demand is reached. We don’t mean to be flippant but it doesn’t make that much of a difference if peak oil demand is reached in 2030 or earlier. And if oil demand doesn’t plateau for several years, but declines. The key is that peak oil demand discussion has overlooked the most basic of oil industry concepts – the oil industry has to replace declines every year on the reality of oil production declines every year. And with decline rates is the challenge to add new oil supply to firstly replace annual declines and then provide for growth up until peak oil demand. Core Laboratories estimates the global world wide annual decline rate in oil is 3.3%. Based on current oil supply of ~97 million b/d, this means the world has to add 3.2 million b/d to stay flat, which then gets to ~4.5 million b/d when you add on growth in demand. If we assume that peak oil demand is reached in 2030 and that the growth in demand slows in the 2020s, oil demand should still be 105 to 110 million b/d in 2030. Decline rates aren’t going to be going down, rather we would expect to see the decline rate higher. But using the current 3.3%, peak demand will still require replacing ~3.5 million b/d every year that demand plateaus. Even if oil declines 1 million b/d per year after 2030, it will still need to add ~2.3 million b/d of new supply every year to keep decline only at 1 million b/d. to go down to 100 million b/d in 2040, and down to 90 million b/d in 2050.
We believe the marginal supply costs are way higher than current long dated strips. We recognize that strips over the near term reflect the dynamics of storage and are used extensively in financial markets. However, we don’t believe long dated strips (2020 thru 2025) are predictive of future prices and do not reflect marginal supply costs. If long dated strips reflected marginal supply costs, they would be inferring that 4 to 5 mmb/d of new supply (to replace declines and support modest growth in demand) can be added every year from 2020 thru 2025 below $52/b. If feels like the narrative on US shale causes a crop market view to oil markets in total ie. high soybean prices one year, farmers plant more soybean, the next year supply goes way up and brings price down. Its one thing to assume that US shale can add another 1 mmb/d below $52, but its another to assume that there is 2 or 3 mmb/d at that price that can be added every year. When we think of the basics of oil supply of having to replace declines, its why we don’t believe the long dated strips are predictive of the future marginal supply cost to add 4 to 5 mmb/d every year up for the next several years. This was also reinforced by WTI being up $1.68 to $57.32, but the long dated strips in 2020 thru 2025 barely moved
WTI Strips Thru 2025
Peak oil demand is going to happen, but the reality of the challenge to add 4 to 5 mmb/d to get there is what will drive oil sooner than expected to $60 plus with a view of longevity thereat. We believe that oil moves to a base of $55 and $60 soon thereafter as the market realizes the oil supply pictures isn’t as good as expected and that the >300 million barrel oversupply to start 2017 was caused by 7 countries adding 7 million b/d of net production from 2014 thru 2016 and that growth won’t happen again at $50 or $55 oil. The key realization being US shale alone can’t grow enough to keep a $50 oil price and even its ability to strongly grow at $50 without a return of equity/debt injections is questionable. Oil then moves quickly to a base of >$60 as the market realizes key fundamentals: peak oil demand isn’t happening as quickly as expected, even once peak oil demand is reached the world will still need to add >3 mmb/d every year to replace declines, and the marginal supply cost for this much oil to be added every year is well over $60. This will be when markets realize there will be some longevity to this call.