If you’ve ever been unfortunate enough to be trapped in a tedious argument about whether or not electric vehicles are genuinely, holistically, and completely better for the environment, then I suppose I either have good news or bad news for you, depending on which side you were arguing on.
The basic question has always been this: are EVs actually better for the environment, even in areas where electricity comes from some decidedly non-environmentally-friendly sources like coal? Are EV drivers, as they can be accused of in these debates, actually driving coal-powered cars, just with more steps? Or are EVs actually as good as they purport to be, and, even when factoring in the environmental tolls of sourcing their battery chemicals and rare earth elements and all that, still better, cumulatively?
The question seems to have at least a bit more of an answer now, as a study published in Environmental Science & Technology’s annual Swimsuit Issue (I’m kidding, they do that bi-monthly) titled Greenhouse Gas Reductions Driven by Vehicle Electrification across Powertrains, Classes, Locations, and Use Patterns, which is a study that, significantly, created a full womb-to-tomb life-cycle model for light duty vehicles (LDV, which I have to admit, is an initialism I’ve never used before, nor wish to use) including cars of varying sizes and body styles and pickups. The model looked at data that factored in a lot of use case variables, like location, climate variations (both seasonal and regional), charging patterns for the EVs, driver types based on usage (commuting, road trips, hauling up to 2,500 pounds of whatever), and so on, all to get as nuanced and complete a view of just how much greenhouse gas emissions these vehicles produced.
Here’s the abstract of the study, from the authors, who let’s shout out right now: Elizabeth Smith, Maxwell Woody, Timothy J. Wallington, Christian Hitt, Hyung Chul Kim, Alan I. Taub, and Gregory A. Keoleian.
“We assess the cradle-to-grave greenhouse gas (GHG) emissions of current (2025) light-duty vehicles (LDV) across powertrains, vehicle classes, and locations. We create driver archetypes (commuters, occasional long-distance travelers, contractors), simulate different use patterns (drive cycles, utility factors, cargo loads) and characterize GHG emissions using an attributional approach. Driven by grid decarbonization and improved electric vehicle efficiency, we are first to report electric vehicles have lower GHG emissions than gasoline vehicles in every county across the contiguous United States. On average, a 300-mile range battery electric vehicle (BEV) has emissions which are 31–36% lower than a 50-mile range plug-in hybrid electric vehicle (PHEV), 63–65% lower than a hybrid electric vehicle (HEV), and 71–73% lower than an internal combustion engine vehicle (ICEV). Downsizing also reduces emissions, with a compact ICEV having 34% lower emissions than an ICEV pickup. We present the first evaluation of LDV emissions while hauling cargo, showing that carrying 2500 lbs. in a pickup increases BEV emissions by 13% (134 to 152 g CO2e/mile) compared to 22% (486 to 592 g CO2e/mile) for an ICEV. Emissions maps and vehicle powertrain/class matrices highlight the interplay between vehicle classes, powertrains, locations, and use patterns, and provide insights for consumers, manufacturers, and policymakers.”
In case you’re one of those strange people who doesn’t want to slog through every published study they encounter on the internet, I’ll try to hit some of the big highlights for you here, so, you know, spoiler alert. Here’s the big one:
Battery electric vehicles have significantly and consistently less output of greenhouse gas emissions than combustion vehicles or hybrid vehicles (even plug-in ones), even in locations where the electricity comes from filthy, filthy coal. On a county-by-county basis in America, battery-powered EVs outperformed combustion vehicles in every single county. All of them.
If we break this down into percentages, we find that BEVs – they specify the average here are ones that have at least a 300 mile range, which definitely isn’t all EVs, though testing factored in 200, 300, and 400-mile range cars – produced 31 to 36% fewer greenhouse emissions than even a plug-in hybrid car, 63 to 65% fewer (I want to say less here so badly) than a conventional hybrid, and a staggering 71 to 73% less than a gasoline-powered combustion car.
The study also notes that smaller combustion cars produce fewer emissions (duh, and 34% less), towing or hauling 2,500 pounds produces more emissions (duh again, 13% more for EVs, 22% more for combustion cars), and while none of these results are exactly shocking, it’s notable to see them so well-supported as in this study.
When it comes to materials, especially battery materials for EVs, and manufacturing emissions, the study used a specific model, which you can learn more about here. Battery replacements were not factored in:
“Vehicle cycle emissions (materials, manufacturing, and endof-life) are calculated using the GREET 2023 model from Argonne National Laboratory. We modified battery size and curb weight for each of the corresponding Car, SUV and Pickup options using vehicle parameters for model year 2025 (SI Note 2). Vehicle cycle emissions include production of components and fluids over the vehicle lifetime along with assembly and disposal of the vehicle. We do not include Li-ion battery replacements during the vehicle lifetime. The latest data shows that for new models, batteries tend to outlast the vehicle’s useful life. We assumed a battery chemistry of NMC811 for BEV, PHEV, and HEV as an example of a high nickel chemistry, the most common chemistry in the current U.S. EV market. For completeness, the impact of assuming NMC111, NMC622, or LFP battery chemistry is also explored and discussed in SI Note 7. The total emissions for vehicles with these other battery chemistries differ by less than 2.5% from those with NMC811.”
There’s a lot of other interesting details in the study; for example, their “use phase” calculations, which give their lifetime mileage estimates for different vehicle classes: sedans are considered to have a lifespan of 191,386 miles (that feels low to me?), SUVs last for 211,197 miles, and trucks at 244,179 miles. I’m not entirely clear how those estimates were calculated, but it’s interesting to see that the baseline amount of lifetime miles for a car has effectively doubled from the roughly 100,000-mile number that seemed to be the accepted standard of the past.
Here’s the equations used for calculating emissions for those lifetime miles, for combustion cars/hybrids and for BEVs:
Look at that, there’s a big sigma there! That’s some real math going on!
The study didn’t go into anything like ease of charging or frequency or made any assessments about the national charging network, or anything like that. It was undertaken just to get an answer to the question of what sort of vehicle drivetrain produces the fewest greenhouse emissions, and, even after factoring in manufacturing, transport, materials, and where and how electricity is produced, it does appear that battery electric vehicles don’t just produce less emissions at their non-existent tailpipes, but also across the board.
I know bringing up coal plants was a satisfying way to get impossibly smug EV-advocates to maybe shut up for five glorious minutes during an argument, but it looks like we’re all going to just have to let that one go. EVs produce fewer emissions, period, across the board. I think if this knowledge is making you feel uncomfortable in some way, perhaps it’s best to just avoid these sorts of tedious debates.
I know that’s what I’m going to do.
Top photo:
I know this wasn’t germane to their study, but I’m most curious about the effect of BEVs on the roads due to their above average weight compared to ICE vehicles (i.e. wear and tear, frequency of resurfacing and the costs associated with that, etc.).
I’m reasonably certain that road wear increases with the square of the vehicle mass, or at the very least, is nonlinear (ie a 5000lb vehicle produces more than twice the road wear of a 2500lb vehicle). It also depends on stuff like load factor – a heavy vehicle with lots of contact patches like a semi truck, vs vehicles like the new Hummer EV that weigh like 9000lb with only four contact patches.
There’s definitely a discussion to be had about the overall fleet of the US and how vehicle mass may be increasing on average. I don’t know if commercial vehicles are getting heavier, but I’d venture to guess that personal vehicles are. I don’t know what the largest contributor to that is – if it’s EVs or drivers’ preference for trucks, or what combination of both we’re looking at.
Some fun studies looking at road wear and tire particulate emissions though:
https://www.sciencedirect.com/science/article/pii/S2590162121000216 (looks at personal vehicles, study undertaken in the US)
https://www.sciencedirect.com/science/article/pii/S2590162121000216 (commercial vehicles, study from NZ)
If you live near rare element mining site, EV not so great for environment.
Catalytic converters on ICE cars also rely on rare earth minerals.
There’s no free lunch. Even bicycles use extractive materials for their tires. Only walking is mostly emission neutral (not entirely https://time.com/7299648/what-is-fart-walk/ ).
If you live on the planet earth, fossil fuel burning cars, trucks, ships, aircraft, etc not so great for environment.
Kudos to anyone who does the work to put valid numbers to this technical issue, but this study doesn’t fully address the environmental costs of EVs. It’s widely documented that EV tires cost more and wear out quicker. The weight of an extra half-ton of EV battery is the main cause of this.
There’s also an opportunity cost to the large batteries in pure EVs. These figures indicate that EVs have twice the CO2 benefit of PHEVs, but they require 4-8x more battery materials. There’s an environmental cost to mining and refining these elements, and an opportunity cost to devoting them to auto production rather than other industrial or utility applications that have a better utilization factor. The average car sits parked 95% of its life, creating no benefit to anyone.
That’s because the scope of the study is about fuel savings.
You talk next about the cost of mining while conveniently ignoring the (environmental) cost of mining coal, gas and oil. Remember the Exxon Valdez (1989)? The BP Deep Water Horizon spills in the Gulf (2010)? The impact of fracking on aquifers? Again – this study is about fuel savings. EV battery chemistry is changing rapidly, with sodium-ion a very promising one which could remove the need for Lithium.
The Model 3 weight is right between the 4-cylinder Genesis G70 and the 6-cylinder G70.
The weight increase for EVs is often highly exaggerated.
We recently put over 1000 miles on a Camry hybrid rental, averaging a real 47 mpg mostly in the fast lane. Nothing wrong with that, especially given it’s quick acceleration and no need to search out EV charging stations. Camry is much cheaper than most EV’s, so you start way ahead dollar-wise.
Seems like a solid study. Can’t say I’m shocked by the findings, I’ve always figured majority electric is where we are going, just the last few years we did a real bad job with the how we get there, a more gradual approach with aggressive hybridization and then more EREVs followed by EVs (with some EVs always available for those for whom they work) along with consistent expansion of the charging networks would’ve been the way to go. Instead we tried to move too fast and scared people which made the whole concept into an easy political football.
So if we get everyone into mild hybrids and PHEV we can cut emissions by half or more, not strain the power grid, and have vehicles around the same price.
Cool. I’ll take a GMC Express PHEV used in 10 years.
We just got done dispelling one piece of false information, and now you throw in another one, that the grid is strained. Simply not the case. Not everyone will ever be charging at the same time. Ever.
The grid strain is from AI intended data centers.
Can we just not with AI I’m so sick of everying Ai these days such a waste.
Ah, its just the latest tech bro VC cash grab. It will go away with flying cars, quantum computers, and self-driving cars shortly.
It isn’t now (most of the time) and won’t be in the future if we implement good charging standards and time of use.
sedans are considered to have a lifespan of 191,386 miles?
Honestly this seems high. For us Autopians I’m sure we have many vehicles that routinely exceed that, but for a normie? No way!
Those people dump their car at 50-100k and the 2nd owner doesn’t care as much and the BHPH owner doesn’t care at all.
I have 2/4 with over that many, and one at 177k. Crap, my average across all 4 is 174k.
And we are Autopian readers (avg 120k here) I feel like there numbers are too high especially if you factor in cars totaled by accidents too.
This study is looking at emissions for the car’s lifetime. Given an accurate lifetime is a predicate for an accurate outcome I’m sure some care was spent in selecting those numbers.
Shopping used cars makes that number feel low. The number of people who sell cars with 200k miles for several thousand dollars would suggest that the buyers of those cars expect to get a fair amount of use out of them.
Averages are interesting, since the outliers (the very few million mile vehicles and those totaled within the first few thousand miles) feel like they would have significant effect, but are probably not affecting the average that much (and the data may remove the extreme outliers).
My personal experience differs. The lessee for the first three years doesn’t care beyond taking it in at the maintenance interval. The buyer who keeps it 4 years and trades in before it’s paid off isn’t abusing it, but not babying it. But the second buyer of both? That’s a person who wants something relatively new, but isn’t shopping new cars. It’s a person who is going to care for that car, maybe because they’ve been waiting a few years to find one they could afford or maybe because it’s the nicest, newest car they could afford. And the BHPH buyers are kind of split between the people who won’t care and the people who will really care, because that was the only place they could get a car. Also, a lot of those are up around 200k, so we’re talking about extending this average at that point.
FWIW a typical US coal plant as of now(ish) emits between 740-1689 g CO2e/kWh. From that we can see a Tesla 3 that uses 26 kWh/100 mi is at best putting out 19.2kg/100 miles and at worst 43.9kg/100 miles using just coal sourced electricity. This does not include the emissions produced in the mining, transporting coal to the plant, transmission and charging losses, cold weather performance, etc which will increase those numbers.
https://www.cowi.com/news-and-press/news/2023/comparing-co2-emissions-from-different-energy-sources/
A Corolla OTOH uses 2.9 gallons of regular to go 100 miles and the hybrid uses 2.1 gallons to go the same distance
https://www.fueleconomy.gov/feg/Find.do?action=sbs&id=49154&id=48490&id=48494
If we use 8,887 grams CO2/ gallon that puts the sin of driving a gasser Corolla at 25.7kg/100 miles and a hybrid Corolla at 18.7kg/100 miles.
https://www.epa.gov/greenvehicles/greenhouse-gas-emissions-typical-passenger-vehicle
That does not include the emissions of oil pumping, refining, and distribution.
So by these numbers the hybrid Corolla is the winner likely followed by hybrid Corolla with the Tesla BEV last.
With your math the hybrid Corolla is twice as good as a BEV yet they say the BEV is better. That means oil pumping, refining, and distribution accounts for significantly more half of the emissions of a hybrid.
Or they messed up something in their data, like a sample bias.
“That means oil pumping, refining, and distribution accounts for significantly more half of the emissions of a hybrid.”
I don’t think so:
“I sat down and read the GM/Argonne National Laboratories well to tank and well to wheels studies to figure this out, as part of an effort to do accurate calculations for my own converted EV’s energy and GHG performance.
https://greet.es.anl.gov/publication-wft2tv3v
The GM/ANL study and the resulting GREET model are very complex, as you would expect given that oil refineries produce a lot of different fuels-gasoline, diesel and jet fuel, home heating oil and bunker fuel for ships, propane etc. But refineries produce a host of other products, and also internally recycle a lot of material and energy, often sharing energy and materials across the fence with numerous petrochemical plants which don’t even count as part of the refinery itself, and with the greater electrical grid.
The end result of a very careful and well-documented analysis via the GREET model which takes all of this complexity into account, is that from well to gas tank, gasoline is about 81 to 83% source energy efficient.”
Which means that even if only gasoline were being burned to get the energy need to get from oil well to gas tank you’d only increase the emissions about 17-19%
Compare that to the additional losses getting energy from mine to battery. I can’t find good numbers on how much energy it takes to dig out the coal and transport it to the power plant but I do have some numbers on what happens to the electricity:
“EV charging loss due to the on-board charger:
Sadly, the on-board chargers are the ones to blame the most when it comes to energy loss as they are usually between 75 and 95 percent efficient.”
https://go-e.com/en/magazine/ev-charging-losses
“The U.S. Energy Information Administration (EIA) estimates that annual electricity transmission and distribution (T&D) losses averaged about 5% of the electricity transmitted and distributed in the United States in 2018 through 2022.”
https://www.eia.gov/tools/faqs/faq.php?id=105&t=3
So right there coal loses between 10-30% of the energy between the power plant and the battery.
“Or they messed up something in their data, like a sample bias.”
Maybe. Or perhaps it is I who messed up. I’m not an expert in this field and it’s a complicated topic with lots of devilish detail. But for now my crude envelope numbers aren’t looking good for coal powered EVs.
When driving ICE cars still emit tons of emissions. The EV car does not.
It doesn’t matter if ICE cars are 100%, 50% or 10% energy efficient ; they’re burning fossil fuels which are polluting the world. Our world. Air you like to breathe in, without the need for filters. Air you would like to be be clean so the crops aren’t covered with soot from the trucks driving past it.
First the study shows there are fuel savings. The study is pretty clear about that. Next we can argue where the electricity comes from or should come from. The more people complain that coal is ‘just as dirty’ the more it makes sense to get rid of coal, quickly.
Once we have tons of wind turbines, solar, more tidal and geothermal, then pollution from cars can be reduced to the absolute minimum.
But look at his first post, a hybrid Corolla is emitting less greenhouse gasses per mile then the Tesla, which means less emissions, and if we consider that a coal plant also emits soot into the air there is a good chance it also emits less particulates into the air too. There isn’t anything magical about ICE that means it’s automatically worse for the air.
The problem seems to be coal, it was only good in the past because it was cheap, it’s time to move on.
I am actually very impressed by the researching you’ve done and commend you for it. Have you looked into the efficiencies of gas/diesel powered vehicles however? They’re weird, neither can idle anything close to efficient, and where they become most efficient is so narrow of a load and RPM range yet crazily tunable. You can probably get away with high 30% – 40% in a gas car, 40’s – 50’s in a diesel if they’re trying, but in most cases expect almost half. I feel like it was about 10 years ago that a prius hit 40% fuel efficiency at the same time a hybrid F1 car finally hit 50% thermodynamic efficiency.
Thanks!
“Have you looked into the efficiencies of gas/diesel powered vehicles however? They’re weird, neither can idle anything close to efficient, and where they become most efficient is so narrow of a load and RPM range yet crazily tunable”
I have. I think that cylinder deactivation, hit and miss operation, 2,6,8 stroke cycles are strategies that may allow engines to widen their efficiency zones, a subject I think is ripe for another Autopian deep dive.
“I feel like it was about 10 years ago that a prius hit 40% fuel efficiency at the same time a hybrid F1 car finally hit 50% thermodynamic efficiency.”
The Toyota announcement was about 2018 while the Merc F1 was 2017. I’ve been seeing rumors the Gen6 Prints will hit 53% TE:
https://priuschat.com/threads/gen-6-prius-engine-will-be-a-%E2%80%9Cgame-changer-%E2%80%9D-achieve-a-53-thermal-efficiency.249167/
Will this pan out or be another dissapointing Skyactiv3 nothingburger? I dunno.
Now imagine that electricity is coming from hydro, wind, solar, tidal or geothermal…
Your comment shows we have to get rid of coal, fast.
The nice thing about coal though is it works when the sun isn’t shining, the air is still, where the ground is parched, far from the sea and where the local geology is dull….
…Of course so does natural gas. Better even. And renewably. And of course there’s nuclear.
Any chance the Corolla and 3 aren’t the same size /class of sedan? Did they get that granular?
I don’t think it matters that much.
So, is the question who made a math error and did not get it checked, or do you think they’re leaning on the data too hard? Maybe they’re averaging in all those bs alternator and lead acid “hybrids” with the “real” ones??
I dunno, maybe their idea of ICE is a 1975 Lincoln MkIV and their idea of a hybrid is a Corvette E-Ray.
I have 2 guesses
It might come down to the fact that most real world EV’s right now are optimized for aerodynamic efficiency, while some ICE are and some aren’t.
If you took the exact same car with the 4 different sources of propulsion you would get different numbers than the averages they present.
The other is I think they are overestimating the lifetime miles of a car, it’s hard to get an exact estimate but ~153k miles seems like the average excluding totaled vehicles, they seem to assume future cars will last longer (hah, not with all this new tech), which will give an advantage to the EV.
Hey now, my 3 car fleet average is around 143000 with 2 at or above your #. DON’T JINX IT!!!
Well yes.
But you are putting the worst case for an EV against a pretty good case for the non-EV.
A hybrid corolla is a really efficient HEV.
Now look at https://app.electricitymaps.com/. You will find that only the WAPA Rocky Mountains is at this moment over 740 g CO2/kwh at 741. Every other grid is lower. Even where coal is used a lot. Most plases in US are significantly lower, like ERCOT (Texas), which is at 368 g CO2/kwh.
Once again, the exception proves the rule. Or in this case, the truth in the article.
The challenge was whether an EV that is EXCLUSIVELY powered by coal today is cleaner than an ICE vehicle or hybrid, not powered by a mix of sources. Which is almost the worst case (see below). The choice of hybrid Corolla was because it’s a common enough vehicle currently on the market with an ICE only version for a more honest comparison but there are many hybrids that offer similar performance.
I should also point out this was not quite the worst case as I only compared fair weather performance. The real worst case would be cold weather performance where the battery capacity drops due to the cold and the energy needed for heat.
Meanwhile the ICE efficiency goes up using cogen math. The waste heat generated by ICE becomes advantageous.
Sure one could argue coal plants can use cogen operation too but I don’t think its very common. Most such systems I’m aware of are generators incorporated into large commercial buildings and AFAIK those don’t burn coal but natural gas.
Blah blah blah seems like they selected a best case scenario for the EV and still used scientific lingo to confuse.
You know the old saying I read on a T-shirt while camping on Asategue island that my parents would not let me have “if you can’t dazzle them with brilliance baffle them with bullshit.” But Jason if you state you understand that survey and believe it is 100% accurate I’m onboard.
Uh
Scientific papers aren’t communication, they’re proof of work in a standardized format designed to provide the details and background for other people in that speciality to peer review, criticize, and replicate. Think of it like a set of architectural drawings, a metallurgy report, or any other industry-specific document: packed with specialty jargon and language.
That is (part) of what makes good journalism so important: it acts as a translation layer, providing information and putting it in context. So you do right by finding a journalist you trust.
I’ve avoided the GHG argument by asking which middle east terrorist organizations are supported by the challenger’s gas money. Not really fair, but they stop and think for a second and, especially if they see my veteran license plate, they go away. An EV won’t fit every use case, but there’s no point in either side dunking on the other for their choices.
Saw this bumper sticker: Help defund oil country terrorists. Drive an EV.
The joke’s on everyone as the Saudis and others in that area are investing heavily in EVs and solar tech, because they’re not dumb.