The End of Fossil-Fuelled Cars We all love a David-and-Goliath story. Elon Musk is a great showman and has played this narrative for all it’s worth in the spectacular rise of his electric car company Tesla. A few weeks ago Tesla launched the Model 3, hyped as the first mass-market full-range all-electric car. But what we really want to know is not whether Tesla will beat Ford and GM and VW, and whose shareholders will make or lose money. It is the answers to two questions:
- Is the electric vehicle revolution on track to take over from fossil-fuel vehicles in the coming decades?
- If yes, will the revolution happen fast enough to be consistent with the Paris 2°C warming cap and net zero emissions not too long after 2050?
Here is a growth chart for global sales of plug-in electric vehicles (EVs).
Below is another one from me, using IEA data — copied into a spreadsheet here. The growth rate for pure battery cars is higher than for hybrids, a complex transitional technology that is fated to fade out.
The rate bounces around and hasn’t settled down, so it would be hazardous to pick a CAGR as a prediction. You can pick a rate and find an expert who predicts it. A nice chart from BNEF shows the scatter.
BNEF leaves out the outlier, Stanford professor Tony Seba, who predicts EVs will be 100% of new sales by 2030. His growth rate is not particularly high – 33% CAGR by my calculations. He does expect self-driving plus ridesharing to slash the size of the market by two-thirds.
I see no reason not to join the party with my own blunderbuss. The growth rate looks to be higher than the 42% that gives a doubling time of 2 years. The global light vehicle market is currently 93.5 million a year, growing at 1.5%. Project the 42% CAGR, and EV sales take all the market in 2031, in line with Seba (see my spreadsheet, and footnote). That’s without his ridesharing revolution.
In all the professional scenarios, the answer to the first question is yes. Electric cars will take over. So will buses and light commercial vehicles (sooner), and eventually heavy ones.
But the answer to the second question is still very doubtful. The problem is that the stock of ICEVs is huge, over one billion, and they last a long time, about 20 years for cars. Even if no new ICEVs are sold after 2030, emissions from the fleet don’t hit zero until 2050. Ridesharing does not affect this problem unless it accelerates scrapping of the legacy stock. Net zero by 2050 requires extremely high growth rates to hold up; follow anybody else’s forecasts than Seba’s and mine, and the deadline is put off. The governments of the UK and France have announced ICE cutoffs for cars in 2040: this is not too bad, as cars over 10 years old tend to have low annual mileage. Net zero by 2050 also calls for extension of the technology to trucks. This is all possible, but not at all certain.
It’s easier to argue that this can happen than that it will. I can at least offer you a qualitative list of downside and upside risk factors.
Downside risks
1. Unforeseen technology roadblocks
These are always possible, but extremely unlikely. The few cases I can think of where technology has stalled – including supersonic passenger flight and nuclear power generation – arose from cost and complexity. EVs, like solar panels and wind turbines, are fairly simple and individually cheap devices that can be mass-produced, the ideal case for continued improvement. As with solar panels, current technology has reached the point where the performance of leading models (Teslas are the benchmark) is already adequate for most purposes, except in the heavy truck segment. The transition would continue solely on the basis of cost reductions.
2. Running out of raw materials
A lithium-ion battery uses lithium, obviously. Depending on the design, it can also require iron, sulphur, manganese, aluminium, nickel, and cobalt. Most of these are abundant, and resource worries focus on lithium and cobalt. They are not very common elements, but neither are they truly rare, like gold and platinum.
Wikipedia gives the abundance in the crust and annual production of the elements. I’ve calculated the mining intensity of a few.
The difference between lithium and cobalt on the one hand, and tin and lead on the other, is that the latter have been sought after for at least 2,500 years and the former for just a few decades. Unless lithium is freakishly distributed, and the few mining operations suggest that it is not, there is plenty of scope for expansion. If lithium and cobalt were mined at the same intensity as the slightly less common lead, output would be respectively 212 times and 82 times higher than today. EV light vehicles have to grow 133 times to take all the car market, so the order of magnitude looks doable. It is also reasonable to expect that raw materials will be used more efficiently as time passes.
3. Policy reversals
Unlike the first two, these are extremely likely. Growth to date has depended on large subsidies, tax breaks, and regulatory perks. In the USA, the federal income tax credit of $7,500 is topped up by a $2,500 rebate in California and a $5,000 tax credit in Colorado. In Ontario, the rebate is C$14,000 rebate plus carpool lane access. In France, it’s a rebate of up to €10,000 or 20% of the car’s price. In Norway, electric cars are exempted from purchase tax, VAT, and highway tolls. In China, the subsidies are backed up with the usual big stick. All these are constantly being changed and will go in time. The risk is that they will be withdrawn before EVs hit unsubsidised price parity. The US federal tax break expires for each manufacturer after domestic EV sales pass 200,000, which will soon be the case for Tesla.
It’s a fair risk in some countries: but not overall. Several countries like Germany are more likely to increase incentives than cut them. More important, in contrast to the early days of mass solar PV, which was completely dependent on the German market, EV sales are roughly equally distributed between the USA, a handful of European countries, and China. (Though, the US market should really be split into California and “the rest of the USA.”)
This robust dispersion provides insurance against a reversal in one market. The Kochs are cranking up a populist campaign in the USA against EVs, but it’s feeble and in any case has no chance of working in California. Pseudoscience on air pollution does not work against the entire US medical profession. In addition, the EV industry is now large enough to pay its own lobbyists: the established automakers are dithering, and the EVs they sell are still a sideline, but they agree that the future is electric, just not yet. They will oppose the Kochs. The well-connected electric utilities love EVs, as they charge mainly off-peak at night, with the future bonus of flexible load management: extra demand with little increase in capacity.
4. A crash in the oil price
The cost equation for EVs depends on cheap electricity compensating for a higher purchase price. If the oil price falls a lot, EVs become less attractive. It’s hard to see how this can become a major downside. Peak cheap oil has already long passed; supply is maintained from more expensive sources, fracking, tar sands, offshore provinces, and secondary recovery. Any further fall in price should be self-limiting from the fall in supply.
It’s more likely that the oil price will spike at least once more before the industry faces its Fimbulwinter. Demand from vehicles is still growing, and investment in exploration and development has been cut following the fall in price. Its immediate cause, the US fracking boom, is necessarily short-lived. A price spike would accelerate the growth of EVs. It’s not so obvious it would trigger another oil investment boom, as by then everybody in the industry will see the end coming. Permanently high oil prices are a possibility. Oil risks are on balance an upside for EVs.
Against these limited downsides, the upside risks are both probable and extensive.
5. Incremental technology improvements
The crucial factor in the growth of electric cars has been lower prices for batteries, from economies of scale and tweaks to technology. Since much of the car is the same in an ICEV (body, seating, wheels, lights, steering, heating/cooling, entertainment) and the costs of these components are similar for all types of car in a market segment, it is all down to the powertrain. Below is a chart of estimated battery cost up to last year (this year’s price is lower still), but note that it is very much based on estimates – it’s an oligopolistic market and the actual contracts hammered out between a handful of sellers at scale and a handful of buyers are commercial secrets.
The quality of electric cars has improved enormously. The first iPhone in 2007 had all the functions of today’s except the fingerprint reader: phone, touchscreen, browser, music player, app store, wifi, camera. All of these have got better and faster, but it’s the same device really. Similarly, a solar panel of 1997 did the same job as one of 2017 in the same way, just at somewhat lower efficiency. EVs have gained much more, in range and charging speed. Here are some all-electric ranges:
If you bought a BEV in 2007, you normally had to sacrifice range and accept the need to plan long trips like a hike in the wilderness. Now you don’t.
It seems likely that these improvements in quality and cost are already built into the trend growth rate, so continuance will not necessarily accelerate it. Indeed, range is subject to diminishing returns. A jump from 100 miles to 200 is significant. From 200 to 400 miles, barely. However, the selection of models and styles on offer will certainly increase. No large company yet is offering an electric pickup to meet this strange fad that Americans share with Third World rebels, though you can get one from niche manufacturer Workhorse.
6. New markets
Commercial and utility vehicles are more important for carbon and other emissions than their raw numbers suggest, because they typically do high mileages, often in urban areas. The fleet therefore turns over faster than cars, which is important for the emissions pathway.
- Electric buses are already available, with adequate daily range of 150 miles and a claimed total cost of ownership competitive with diesels. They have a fifth of the Chinese market, 100,000 sales a year. I expect rapid change in new buys by city operators.
- Light vans (universally used in Europe in place of pickups) are coming onto the market now. The German Post Office designed its own after being rebuffed by VW, and has now gone into partnership with Ford Europe to sell it. Renault and VW are bringing out 3-ton vans (Transit type) in Europe this autumn, with ranges around 100 miles, plenty for urban cycles. Renault and Nissan already sell the small Kangoo ZE van and larger e-NV200.
- Even heavy, long-distance trucks are on the way. Tesla claims it will demonstrate an all-electric US semi in September. Volvo, MAN, Eaton, and Mercedes are all working on hybrid powertrains. “Hybrid” here can mean anything from “a few electric miles in cities at either end of the run” to “an emergency diesel range extender for a basically electric vehicle.” One issue for BEV trucks is that their enormous batteries will need massive chargers, at least 350 KW, in a comprehensive network requiring some local grid upgrades.
7. Getting really cheap
The current policy incentives are designed roughly to equalise costs, not undercut fossil fuel cars. But as we saw with wind and solar electricity (and as I predicted), the price of EVs won’t stop falling once it reaches parity with ICEVs. After parity, the widening price gap will steadily boost EV sales; and as it will be driven by economies of scale, it’s a positive feedback loop – or death spiral for fossil vehicles. This looks highly probable to me. The price drop will not be as spectacular as for solar panels, however, because of the half of the car that does not change.
8. City policies
In the wake of the Paris Agreement, hundreds of cities across the world have signed up to green clubs and adopted more or less vague and ambitious targets for renewable energy or emissions. They have yet to agree on a rigorous methodology for benchmarking, or systematic peer review. As always, some are serious and others no doubt in it for the photo-op. Among the serious is London, which will introduce an Ultra-Low-Emission Zone in September 2020. Vehicles not meeting a high emission standard will have to pay a daily charge of £12.50, collected in the same way as the existing congestion charge. This policy creates a solid policy pressure on fleet operators of vehicles to go electric earlier. The public transport operator, Transport for London, is already electrifying its single-decker fleet and trialling double-deckers.
Even lesser policies can be effective. Stuttgart is considering bans on diesels triggered by air pollution indicators. If you are the fleet manager of a supermarket chain relying on daily delivery, the prospect of even a few days a year of unpredictable interruptions is very worrying.
How many other cities will join London in adopting effective pro-electric policies? My guess is: enough to make a difference. Sadiq Khan is not suffering from the policy – he is much more popular than Donald Trump.
9. Technology breakthroughs
The workhorse lithium-ion battery, invented by Nobel non-prizewinner John Goodenough, does the job. Many other battery chemistries are being investigated, and within each chemistry new nanomaterials for anodes, cathodes, and membranes. A sudden jump in battery performance or fall in cost is entirely possible and would accelerate change. It’s electric aircraft that really need something new.
10. Grid batteries
Once there was a bright idea: once you have a lot of EVs, there is a massive pool of battery capacity you can in principle use to balance and stabilise the electric grid. Suppose you have a Tesla with a 300 mile range, but you know that tomorrow your trip will be at most 50 miles. Then you can sell the spare kilowatt-hours to the grid, as long as you can recharge at night. The first studies threw cold water on this, because the additional charge and discharge cycles shorten the life of the expensive battery. However, Warwick University researchers have found that with good software and careful optimisation, grid use actually extends the life of the battery. They used real cars in the university car park. Warwick sits at the centre of the UK car industry and Motorsport Valley. The work was commissioned by Nissan, which has sold more EVs than anybody else. This looks credible. The vehicle-to-grid (V2G) scheme is back on the agenda. If it takes off, it will provide EV owners with revenue, and lower the net cost of EV ownership. This should speed up sales.
11. Self-driving
There is a lot of hype about the potential for fully autonomous cars, coupled with sharing, to reduce the number of vehicles needed. A fleet of robotaxis could on paper replace 10 times as many privately owned cars. If this pans out, as Tony Seba predicts, the existing ICE fleet would become obsolete much sooner than in the conventional scenario, and the total number of new EVs needed for the transition would be slashed. Weeell. The technical and regulatory challenges for full autonomy are very severe: Google’s test cars navigate Mountain View safely, but that’s a long way from traffic in Delhi or Bangkok. Still, serious money and seriously clever software engineers are working on this dream, and success must be at least an evens bet.
12. Social pressures
The emotive and social signalling component of car purchase and car use is very strong. Elon Musk made a good decision in starting with an impractical sports car, offering stunning acceleration off the lights, designed for would-be celebrities to pick pretty girls or boys up on Rodeo Drive rather than a worthy vehicle with hairshirt connotations. He must be delighted at the idiots who strip current Tesla saloons down for drag racing. We don’t know how these social effects will work out. Still, there is a real possibility that gasoline cars and a fortiori diesels will come to be looked down on as uncool and antisocial, like smoking or spitting in the street.
13. A unicorn: the carbon tax
I include this for completeness – but clearly the policy tool recommended by most economists is a political non-starter for now. It may happen once it’s too late to matter. The Rube Goldberg contraptions we have – tax breaks, city bans, perks, regulations, research subsidies – are very fortunately good enough to do the job.
It seems pretty clear that there are more ways, and more plausible ones, for the electric transport revolution to speed up than to slow down. A pity that the carriage lamps won’t come back.
Footnote 1: The exponential curve
It won’t of course hold to saturation. The standard S-shaped logistic curve implies a slowdown as you near it. But for the last few percent, the exact date is immaterial. There is still far too much uncertainty about the trends for the next decade to worry much about the endgame.
Footnote 2: Terminology
If you have not been following this, there are four categories of more or less electric vehicles:
1. Closed hybrids where the battery takes over for short stretches in towns, and is recharged or paralleled by an internal combustion engine (ICE). Example: old Toyota Prius. Known as plain “hybrids.”
2. Plug-in hybrids with a battery that can be recharged from the grid, known as PHEVs. Example: GM Volt. The all-electric range goes from a token 15 miles to 50 or more.
3. All-battery vehicles, or BEVs. Examples: Nissan LEAF, GM Bolt, any Tesla.
4. Vehicles where the battery is recharged from an on-board fuel cell, FCEVs. Solitary example: Toyota Mirai.
The first and fourth categories are dying. Plain hybrids get the mechanical complication of a double powertrain, but offer no path to zero emissions. Hydrogen fuel cells require a very expensive brand-new fuel distribution system which no private investor is willing to build, and lose efficiency from the double conversion. There is an endian dispute whether PHEVs should count as EVs. I will leave them in, though they only represent a transition technology until batteries are cheaper all round for adequate range.