The 21st century has seen a significant slowdown in the growth of oil consumption, yet robust growth in electricity generation. The divergence will widen further as large pieces of the global economy are electrified, and oil growth falls to zero. The peak in global oil demand was very likely set in 2019. While 2021 will see a major rebound for oil consumption, all years going forward will see oil demand oscillate around a flatline. The IEA (Paris) most recently forecast, for example, that global gasoline demand has now peaked. That’s a decent proxy for the scant growth that lays ahead for oil. One cautionary note: while oil growth is over, the decline of oil consumption is not yet near. That’s due to a combination of factors, in everything from the lengthening lifespan of the existing ICE fleet, to the continued growth of oil demand in the Non-OECD. To be sure, the more pronounced oil decline will eventually come, but probably not until after 2025.
As the EV adoption curve steepens, a common (and wrong) intuition will take hold: the world will not be able to power EV without lots of new coal, nuclear, or natural gas fired power generation. The intuition endures from a lack of understanding. Powering EV represents a massive reduction in energy demand compared to oil, which loses most of its power to waste heat. A typical EV requires at least 60% if not 70% less energy to travel one mile. That’s how you can place 1.367 million new EV on the road, as China did last year, while triggering a fairly light call on new power—just 4.785 TWh of electricity demand.
But note how much new electricity China created from new deployment of wind and solar, also in the same year: 186 TWh. Indeed, over the past several years, China could have placed at least 20X more EV on the road, all covered by new clean electricity from wind and solar. To those who maintain this equation will not last, and that eventually EV demand will swamp powergrids, consider that the cost to cover new power demand from EV is already cheap, is already fast, and just keeps getting cheaper and faster.
One measure of wind and solar’s impressive speed is the rate at which they’re catching up to and surpassing incumbent energy. The contrast with global nuclear is most stark, due to maddeningly long construction timelines for nuclear, as new wind and solar propagate easily. Yes, in a political economy, the yawning gap between nuclear and wind+solar’s deployment rate is influenced by policy choices. But only partly. Wind and solar have intrinsic qualities in their engineering, as does nuclear, that makes bridging that gap all but impossible. This is a case-closed situation: wind and solar have fully and irrevocably won the cost argument against nuclear.
That said, The Gregor Letter continues to advocate that we should build some new nuclear. And, should small modular reactor technology (SMR) eventually get off the ground, instead of serially being “the next little thing,” that would be great news. Until then, however, wind and solar (and now, storage too) are displacing new fossil fuel demand at an effective rate. And there’s a compounding effect associated with that success. So dangling the possibility of new nuclear technology as an eternal promise is frankly unimpressive, when a tonne of carbon avoided in 2021 is a far bigger tonne in compounding terms, than a tonne removed in 2031. And just to add, there is no prospect at all, right now, of further aggregate growth in global nuclear capacity. The oft-touted “50 new reactors under development” fails to cite the fine print: staggered construction timelines stretching out 20 years or more, and at least as challenging, aggregate retirements that will neutralize much if not all of that new capacity.
Sometime later this year or early next, therefore, generation from wind and solar will surpass nuclear globally. In the projection below, nuclear generation is held flat, and wind and solar grow by a very conservative 300 TWh per year. That’s a linear projection of course, and they will surely grow faster. But in short term projections one must account for single year outcomes that temporarily slow the buildout.
You are reading a free issue of The Gregor Letter, a bi-weekly publication that enjoys broad subscribership internationally among academic, financial, and policymaking institutions. The letter is data focused, and relies on industry knowledge gained through regular reporting, and consultation with a global network of experts. The publication cycle is every other Monday, and adheres to a strict deadline: just after midnight Pacific, in time for the morning readership in London.
Subscription rate options are $75 per year or $7.50 per month for individuals, and $300 per year for institutions. For further assistance in arranging a license for institutional subscribers, please see information here on invoicing and subscribing.
Last year’s oil demand crash, the largest ever, gave society an opportunity to think about life without cars. Skies cleared to levels not seen in decades and there was no better city to track that stunning change than Los Angeles. If you examine the bottom row in this EPA chart, note the horizontal bands of green starting in March of last year.
And yet. While California has done a terrific job incentivizing EV sales, preferencing EV drivers, and penalizing ICE cars with extra fees and taxes, this has still not resulted in the decline of state petrol consumption. Importantly, even as road fuel consumption is no longer growing (an outcome the state had long since modeled when projecting the end of gas tax revenues) there is still no plan in any of the state’s three largest conurbations to introduce road and congestion fees. Note the color change back to something more normal in reds and purples, starting in September of 2020.
Thus, in California, we have a fine portrait of a deeply progressive state legislature, deep blue voters in some of America’s largest cities, and elected leaders vocally committed to climate change—all acting together to avoid the hard task: a confrontation with the existing fleet of ICE cars. California remains an excellent proxy for the world as we adopt EV. While future road fuel demand growth is cut off, there is no imminent decline of oil consumption.
The earth contains myriad clocks. Those used in radiometric dating, for example, are colloquially referred to as clocks in rocks. Researchers and historians also use methods to cobble together long term data sets, cleverly using proxy units of account to measure trends. To chart the dramatic rise in wood prices in pre-industrial England, for example, the Oxford scholar of energy and the industrial revolution Robert Allen used silver to track wood’s impressive surge—right before England turned to coal. Reaching back further in time, researchers Yasuyuki Aono and Keiko Kazui have stitched together a chart of Japan’s cherry blossoms back to circa 800 AD. As the Economist magazine explained, temperatures in winter generally determine the date of full-blossoming. More important, the cherry blossom pattern has predictive value:
From its most recent peak in 1829, when full bloom could be expected to come on April 18th, the typical full-flowering date has drifted earlier and earlier. Since 1970, it has usually landed on April 7th. The cause is little mystery. In deciding when to show their shoots, cherry trees rely on temperatures in February and March. Yasuyuki Aono and Keiko Kazui, two Japanese scientists, have demonstrated that the full-blossom date for Kyoto’s cherry trees can predict March temperatures to within 0.1°C. A warmer planet makes for warmer Marches.
The most important pattern in this chart is unmissable: the pronounced change that occurs starting with the industrial revolution in the early 1800’s.
Rebound years are instructive, as they provide useful information about the system we’re currently running—not the one we’ll have in the future. For example, global oil consumption will rise by at least 5 million barrels a day this year, from the deep lows of 2020. That won’t be enough to resurrect global oil demand back to 2019’s highs. But the comeback, a 6% advance, illustrates the composition of the global vehicle fleet.
In a recent paper, Fossil CO2 emissions in the post-covid era, Glen Peters et al. wrote, “Although the measures to tackle the COVID-19 pandemic will reduce emissions by about 7% in 2020, they will not, on their own, cause lasting decreases in emissions because these temporary measures have little impact on the fossil fuel-based infrastructure that sustains the world economy.” One area the paper explores is that high income economies, while successfully restraining their emissions, have done so at very elevated levels. Lower income economies by contrast produce far fewer emissions on a comparative basis, but are still growing. This is yet another lens through which to see the current emissions problem in the West: CO2 growth has been successfully countered, but action to force outright declines is lacking.
Oil consumption on a per capita basis remains far higher in the OECD, than in the Non-OECD. But from a lower base, and with far larger populations in the aggregate, the Non-OECD took control over global demand earlier last decade. The divergence helps to explain the peak gasoline forecast from the IEA. OECD road fuel consumption may be ready to enter very gentle decline. But not in the Non-OECD. Just to note: IEA doesn’t see peak oil demand globally, just yet. Air travel and petrochemicals figure prominently into their reasoning. The Gregor Letter view: global oil demand has peaked, and will bump along a plateau for a while yet.
There is great promise however that Non-OECD oil demand growth too will slow, sometime this decade. The prime mover in this equation is China, of course. In 2017, China halted the growth of ICE sales at 28.102 million units, as EV took control at the margin. By last year, ICE sales had fallen to 23.944 million units. This year, ICE sales should stabilize temporarily around 24 million units, as EV sales hit 2 million.
After a 7 year plateau, global coal consumption is finally ready to enter outright decline. Coal consumption was hammered last year of course, along with other fossil fuels. But underneath the surface, that decline was mostly composed of ongoing retirements in the OECD, even as China’s coal consumption rose again. China is of course already in deep coal overcapacity. The country doesn’t need to build a single new coal plant, but does so regardless, as a component of industrial policy. This will make the collapse all the more spectacular when it arrives. That said, the lengthy plateau of coal’s dependency, from 2013-2020, serves as an ongoing cautionary to how difficult it can be to kick large systems off of their deeply embedded dependency.
—Gregor Macdonald, editor of The Gregor Letter, and Gregor.us
The Gregor Letter is a companion to TerraJoule Publishing, whose current release is Oil Fall. If you've not had a chance to read the Oil Fall series, the single title just published in December and you are strongly encouraged to read it. Just hit the picture below.