Electric Cars

2 June 2016

Like many people, I was amazed and happy to hear that Tesla has plans to make 500,000 cars a year by 2020. They might even be able to do it. Tesla has been known for missing deadlines in the past, but as a software engineer I can forgive them. Even if they only end up hitting the 500K goal by say 2022 or 2023, it will be an impressive feat.

One reason to be excited about electric cars is that they emit no carbon dioxide. True, their power comes from whatever the grid supplies, but as of 2015 the U.S. is down to getting only 33% of its power from coal, with another 33% coming from natural gas and the remaining third coming from greenhouse gas free sources: nuclear (20%), hydro (7%), and renewables (~7%). We're doing far better than even a few years ago, when coal supplied 44% of power in the U.S., and I'm optimistic for the future. We would need to add more power sources to the grid to accomodate electric cars, perhaps twice as much as we currently have, but that seems feasible.

Not to go on too much of a tangent about power, but I consider natural gas to be better than coal for two reasons. The first is that when burned, natural gas produces about half the greenhouse gases as coal. The second is that a ton of coal produces about ~250 pounds of coal ash, which has to be dealt with. (I'm getting the ~250 pounds number based on how in 2012 we burned about 800 million tons of coal, producing nearly 110 million tons of ash.) Of course, neither coal mining nor fracking for natural gas are particularly desirable.

Tangents about fossil fuels aside, the real question that I've been pondering is, if the U.S. switched to only electric cars, how long would it realistically take? But also, how much would it reduce greenhouse gas emissions (assuming the cars were powered from greenhouse gas free power sources)? The answer to the second question is that 26% of greenhouse gas emissions come from transportation in general, of which "over half" is produced by cars, light trucks, SUVs, and the like. (That last link also helpfully points out that 11% of our emissions are absorbed every year by trees and other plants growing in the U.S.)

So let's say that 15% of U.S. greenhouse gas emissions come from small vehicles, and could be eliminated if we all switched over to electric cars (and used emission-free power sources). That's...not as impressive a number as I would have thought. But it is still quite significant--every bit helps. So the next question is, how long would such a changeover take? The answer as I see it has two parts. The first is how long it will take Tesla and other car companies to ramp up their production (assuming there is continual demand for electric cars). The second is, once factory capacity is available, how long it would take Americans to replace say 90% of their vehicles with electric cars. That's assuming that us Americans have sufficient aggregate desire to buy electric cars.

It was hard for me to find a source for how long it takes to build a car factory. Based on one example, it sounds like 3 years is plausible. I was able to find a list of the top producing car factories in the U.S., which seems to indicate that 250,000 cars/year is near the high end. So Tesla will likely need at least 2 factories to make its promised 500,000 cars a year. It might be able to finish them by 2020, but given that the type of car is a new one, that might be a bit optimistic.

Pretty much everything I've given above has been fact, or as close to it as I can get. Now it's time for the fun part, speculation. Let's suppose that Tesla makes it to producing 500K cars by 2021, and that every 3 years thereafter until 2030, it's able to build 2 new factories (which would require significant financing). Then by 2030 it would have a total of 8 factories producing 2 million electric cars a year. Let's assume also that demand picks up for electric vehicles (which may be far-fetched) to the point that other car companies decide to manufacture them in significant numbers, such that along with the 2M cars Tesla is producing, the other car manufacturers are making another 2M electric cars by 2030. So 4M cars/year total. How much of an impact would that be making?

The average age of an American car is now 11.5 years. This implies a lifespan of 23 years (11.5 * 2) for most American cars. There are about 250 million cars in the U.S., implying a turnover of about 10 million a year. So if 4 million new cars in 2030 are electric, that's pretty good, right? Well yes, but because of the 23 year lifespan of cars in the U.S., it would take at least a couple decades to phase out the gasoline burners, and probably more.

Here's some more speculation. Assume 10 million electric cars are made (and purchased) between now and 2030. Further, suppose after 2030 no new electric car factories are built, but people continue to buy the 4 million electric cars produced each year. Then by 2040 there would be about 50 million electric cars on the road, out of 250 million total, or about 20%. The number would grow to 90 million electric cars (almost 40%) by 2050, at which point it would be maxed out.

In a more optimistic scenario, suppose the car companies were able to double their electric car capacity from 4M/year in 2030 to 8M/year by 2040. Then on average maybe 6M/year would be made between 2030 and 2040, which together with the assumed 10M made by 2030 would give 70M electric cars on the road by 2040. If the car companies stop increasing capacity there but produce another 8M/year from 2040 to 2050, that's a total of 150M electric cars on the road (again out of 250M total) by 2050. So about 60% of cars on the road would be electric by then, with about 80% by 2060.

I find these numbers both exhilarating and depressing. On the one hand it's a plausible path to having a majority of the vehicles on the road be electric. On the other hand, on what seems like a plausible but maybe optimistic scenario, it's going to be 35 years before 60% of cars are electric. Maybe increasing and better automation will help matters. Maybe Tesla will be able to build more efficient factories, and build the factories themselves faster. But I would guess that the existing car companies have gotten pretty good at building theirs, so I wonder how much room for improvement Tesla has.

The other question is how willing people will be to buy electric cars. Sure, Tesla got a huge number of preorders for its Model 3, but would a majority of Americans be willing to buy them? My suspicion is that oil prices will keep spiking up and down over the next couple decades. I also suspect that if we can keep increasing our use of renewables in this country along with improved storage technologies, the cost of electric power will be appealingly stable for car buyers. But I guess we'll find out. At least there is a path to reducing emissions from vehicles, but it's going to take a long time, maybe longer than people are expecting. The analysis I described above certainly surprised me the first time I tried out different scenarios. That's why I wrote it up.

home