Gas versus electric: the final showdown
Many of you have been waiting for the straight scoop on the true cost of owning an electric versus a gas car. Well wait no longer, this covers it *ALL.* (You’ve been warned.)
I read an interesting post on the backfire effect the other day and the time has come to change my ‘evil’ ways. No more trying to convince people of things based solely on facts. Hmmm…
The backfire effect: defined as the effect in which “corrections actually increase misperceptions among the group in question.” 
While I admit I begin this post as an electric vehicle enthusiast, I may not be at the end — who knows? I will strive for a cost analysis as uniformly fair as possible, and I very much encourage input from both sides — combustion car enthusiasts and electric vehicle people alike — on mistakes I may have made or anything I may have left out.
I’ll also take advice from a 1/1/17 post in Scientific American by Michael Shermer :
1. keep emotions out of the exchange
2. discuss, don’t attack
3. try to articulate the other position accurately
4. show respect
5. acknowledge that you understand both sides
6. try to show how changing facts does not necessarily mean changing worldviews.
*Please restrict comments to the same.*
With that as a preface, let’s look at the question of which costs more to own and operate — an internal combustion engine (ICE) car, or an electric vehicle (EV)?
For the purposes of this discussion, we’ll set aside the complexity of insurance costs. Consider the below analyses to be a starting point from which you can compare your individual insurance rates between the vehicles you prefer to determine which would likely save you the most money. (In a later post, I’ll cover why the car insurance business is likely to change dramatically in the next ten years or so.)
Aside from insurance costs, we’ll also set aside the daunting prospect of in-depth analyses of used cars — we’ll consider new cars only, and we’ll look at them after three points:
1. six years — which, according to this article  cites research by R.L. Polk as the average length of time drivers keep a new vehicle
2. eleven years — which, according to Consumer Reports  is the average age of all cars on the road
3. fifteen years — which, at the average American driving pace of 13,474 miles/year, amounts to 202,110 miles: slightly better than the 200,000-mile landmark that only 1.3% of cars reach .
Clearly, if you spend enough money on the repairs or maintenance of any vehicle, it may last longer than this. Here I document reasonable repair costs whenever possible (generally possible, at least with combustion cars.) I’ll estimate somewhat high for EV repair costs due to lack of sufficient historical data to reference. I hope this will allow for a reasonably objective comparison.
The average price of a new car according to Edmunds [referenced in 6] was $36,115 as of January 2018. From here I take a different approach than the very good general analysis provided by Two Bit da Vinci on YouTube here  — where they examined just three vehicles of widely varying prices. For this comparison, I will pit a group of three different electric vehicles each in the above price range to a group of five different ICE vehicles in that same price range and five more popular cars at various other price points.
Finally, I will be comparing these vehicles across the range of average state electricity versus gas prices as well as projecting those prices over the next six, eleven, and fifteen years based on a ten-year moving average historical change in the national prices for those commodities — which, to be frank, is one of the most critical components of the analysis. As you can see here , the variability in gas prices, while it trends upward — is hardly a stable figure relative to inflation. [Note: the figures show a ten-year moving average increase of gas prices ~4.4% per year versus ~1.6% per year for electricity. This clearly makes a big difference for projections moving forward, but it’s also understandable for the reasons covered below.] Later (toward the bottom of this post), I will use this national average annual increase in conjunction with available state prices to show you how to estimate the expected six, eleven, and fifteen year gas and electric prices in your state — so you can use this information to determine beyond reasonable doubt which type of car would be least expensive for you to own and operate.
The wide disparity between the historic increase in the price of gasoline versus electricity is obviously based primarily on the fact that the former is nonrenewable, while the latter can be sourced renewably — as well as the fact that electricity’s price is rendered more stable by the myriad ways in which it can be generated. Projecting these numbers out to their expected six, eleven, and fifteen-year levels yields the following:
Gas price, national average:
· six years: 29.5% higher than projected 2018 levels (rising to a national average of $3.392/gal in 2024 from $2.62)
· eleven years: 60.6% higher in 2029 (…to $4.207/gal)
· fifteen years: 90.8% higher in 2033 (…to $4.998/gal)
Electricity, (residential) national average:
· six years: 10.1% higher than current levels ($0.1301/kWh to $0.1432/kWh in 2024)
· eleven years: 19.3% higher in 2029 (…to $0.1551/kWh)
· fifteen years: 27.1% higher in 2033 (…to $0.1654/kWh)
I realize this may come as shocking — and perhaps frustrating — news to ICE vehicle enthusiasts, but it should not be at all surprising if you consider in particular the dramatic reductions in solar energy costs we’ve seen in the past ten years. While electricity prices have been remarkably stable over time, it’s easy to forget that gas in the U.S. was hovering at just about $1/gallon twenty years ago, that it rose to over $4/gallon in the ten years which followed, and that last June it averaged $2.42. The price of gas tends to be much more erratic than the price of electricity — which is stabilized by the diversity in modes of generation — but the trend of both over a prolonged period of time is impossible to deny. As mentioned, for gasoline, that trend has been 4.4% per annum over the past twenty years, for electricity, 1.6%.
We should also consider this doesn’t reflect the recently proposed $0.25/gallon Federal gas tax hike which, if passed, stands to tick the price of gas even higher (the last time the gas tax was raised was in 1993.) Also, significant price reductions continue to be expected in grid-scale photovoltaic energy production in part due to the falling prices of panels, and this quickly deployed and inexpensive form of electricity production — coupled with solar’s rapidly increasing market share — will likely dampen electrical price increases into at least the next decade. [Note: for people who install solar on their homes, taking advantage of the sun as a means to power your car can drop the effective fueling cost of an EV far below the numbers estimated above. This obviously means that if you’re interested in saving the most money, you should consider the rest of the analysis relative to what it would be with home solar.] Moving on to the analysis, then…
I’ve taken representative internal combustion (ICE) vehicles from five of the largest manufacturers in the $36K price range:
1. Toyota: 2018 Avalon XLE Premium — ranked #1 in Large Cars by U.S. News and World Report (USNWR) 
2. Volkswagen: 2018 Volkswagen Golf R  — a performance hatchback Edmunds rated at 4.5/5 in 2017 (the base Golf received 8.5/10 from USNWR, and ranked #6 in Compact Cars)
3. Ford: 2018 Fusion Energi Titanium — a plug-in hybrid  with a 7.6 kWh battery (20 miles’ worth of usable range) and a 14 gallon fuel tank. It received 7.6/10 by USNWR, which ranked it #19 in Midsize Cars and #17 in hybrid/electric
4. Honda: 2018 Accord — Kelley Blue Book’s 2018 Best Buy Award Winner , which received 9.3/10 by USNWR and is ranked #2 in Midsize Cars.
5. Cadillac: 2018 Cadillac ATS  — a compact luxury car which was rated 8.3/10 by USNWR, and ranked #11 in Luxury Small Cars.
I’ve also added the most popular cars *excludes trucks* (2017 figures, ) across all price ranges:
- 2018 Toyota RAV4 — a compact SUV which received a 7.7/10 by USNWR, ranked #13 in Compact SUVs.
- 2018 Nissan Rogue — a compact SUV which received a 7.9/10 by USNWR, which ranked it #11 in Compact SUVs
- 2018 Toyota Camry — the #1 Midsize Car, rated 9.4/10 by USNWR
- 2018 Honda CRV — the #1 Compact SUV, with a rating of 8.9/10 by USNWR
- 2018 Honda Civic — the #1 Compact Car, with a rating of 8.8/10 by USNWR, and a perennial favorite model in the U.S.
…and used what I considered representative/economical options packages for each of them (generally the base model.)
Finally, I’ve compared them to the three most widely available electrics in the U.S. market in the referenced price range:
1. Tesla: 2018 Model 3 — which wasn’t rated by USNWR at the time of this writing, but received a 4.5/5 from Edmunds, which praised it for “stunning performance among other EVs” and for being “more technologically advanced than rivals” but listed this disadvantage: a sticker which rises significantly with options (e.g. Long Range battery pack, AWD, etc.)
2. GM: 2018 Chevy Bolt — ranked #6 in Compact Cars and #4 in Hybrid/Electric Cars with a rating of 8.8/10 by USNWR
3. Nissan: 2018 Nissan Leaf — ranked #19 in Compact Cars and #15 in Hybrid/Electric Cars with a rating of 8.0/10 by USNWR
I’ve taken a different approach than YouTube’s Two Bit da Vinci in that I’ve largely kept the figures independent of one another, allowing you to more easily see where the strengths and weaknesses of each vehicle lie. This next table shows MSRP/fuel costs for the internal combustion (ICE) vehicles in the line up:
Next, we’ll look at what these cars would cost per year if they required no maintenance to reach ~200K miles (13,474 x 15 = 202,110)
The residual value on all of these cars can reasonably be expected to be about $1500 +/- $1000 at the 15 year mark — with most having depreciated by ~90-95% of their initial value. What this basically means is that the value of any of those cars in 2033 will likely be between about $1000 and $2000 provided they’re still functional. The scrap value of a junk car is roughly $500 — which will be relevant at the end of year 15 if you keep the car until it dies completely. But remember, just 1.3% of cars make it to this milestone — planning for yours to be the exception is, statistically speaking, unwise.
The reason for presenting the numbers this way should be clear: these figures are prior to any maintenance costs and effectively represent the base starting point from which all of the cars can compete based on reliability — and how much their routine maintenance and repairs will cost. You can already see one reason the Civic tends to be so popular: it’s a very utilitarian vehicle with a yearly operating cost far lower than the rest, at least from the standpoint of its sticker and fuel economy. You could spend about $10K more in repairs on it over the 15 years of its life and still handily beat any of the other gas cars except the Camry — and even with the Camry, it would be close.
Because taking good care of a car and driving it until it dies tends to be the most affordable approach to car ownership, these figures — once maintenance is factored in — are the lowest you can reasonably expect a car to cost you, exclusive of insurance.
[Note: the point can be argued that few people pay the MSRP on a new vehicle — and while this is no doubt true, it is also true that few as a percentage pay in cash. This means the effective price paid will usually vary down from the MSRP slightly and then up somewhat due to financing. The latter makes more sense than paying cash when financing can be obtained at a lower rate than returns from investments normally yield. We’ll not delve too deeply into the dealership hornet’s nest here.]
Now let’s look at the costs associated with the electric vehicles in our lineup. I’ll do this three ways:
1. Without considering any potential rebate incentives
2. Considering the full value of the current $7500 Federal rebate ($4000 for the Ford Fusion Energi, which I’ve placed in both lists for comparison.)
3. Considering one-half the value of the current Federal rebate (=$3750 or $2000 in the case of the Fusion Energi)
The reason for looking at half the rebate value lies in the way the phase out of the Federal rebate is designed. More on rebates later…
This next table shows the current MSRPs of the electrics and their typical anticipated fuel economies. [Note: The all-wheel drive Tesla enjoys a fuel economy advantage over the other vehicles due to distributing its power over two drive motors versus one.]
The fuel economy figures listed here are estimates, and are a composite of my real world experiences with full battery cycling in the 2013 Nissan Leaf, the reports I’ve taken from others, user reports on the EPA’s website, and projected losses from meter to motor. Importantly, this information refutes the EPA’s posted estimates, as I will here discuss in detail.
The EPA’s “miles per gallon equivalent” — or MPGe — is a curious metric which arose in 2010 in part due to the difficulty customers were having understanding fuel economy for electric vehicles. Unfortunately, it unfairly weighs the chemical bond energy of a gallon of gasoline (33.7 kWh) against the more highly extractable energy in the form of chemical energy stored in a battery. Here’s where it gets somewhat complicated, and why such a comparison is inherently flawed.
· Combustion engines rarely top 30% thermal efficiency. That is, roughly the most you can expect from a gas-powered engine is mechanical utilization of about 11–12 kWh of the energy present in a gallon of gas. [Note: Toyota introduced a prototype engine in 2014 which bested this level: 38% — which would boost this number to 12.8 kWh if such an engine actually made it into a car.]
· Electric motors, by contrast, generally exceed 80% efficiency, typically averaging 85–90%— with maximum efficiency prototypes exceeding 99% in some cases [ABB, 2017.] Two reasons for this are the significantly better low rpm torque of electric motors and their comparative lack of friction in comparison to combustion engines.
· While a reliable estimate of the fuel economy of a combustion engine is very easy for a user to determine by merely taking the miles traveled since last fill up and dividing it by the number of gallons required to fill the vehicle again (the EPA testing facilities actually use the vehicle’s exhaust gases to enhance precision), the fuel economy of an electric vehicle tends to be very difficult to determine, for the following reasons:
1. Fuel gauges on electric vehicles often read by percentage, not total fuel remaining in kWh — thus it’s difficult to tell how many kWh you start with and how many you end with. They’re also confounded by so-called ‘guessometers’ — which many users treat like fuel gauges and which seek to predict remaining range based in part on recent consumption. [The Tesla Model S and X both allow a user to see Watt-hour consumption per mile, but again this obviously reads from the dashboard, and presumably reflects the battery charge depletion, not the total electricity per mile the car uses to travel as measured from the source which directly supplied it. From a 110 Volt AC outlet in a person’s home this will be one number (passing through the car’s onboard inverter before reaching the battery), from a 220 Volt home charger it will be slightly different, and from a DC Supercharger station it will be third number. You normally will not know exactly what the number is, nor will you know precisely how many kWh it translated to once it arrived at the battery.
2. The percentage readout on EV fuel gauges, when available, reflects percentage of the battery’s usable capacity, which is not the same as the vehicle’s battery size (the usable capacity is always marginally lower than the total battery capacity.) Thus a reading of 50% from the dashboard of a 24 kWh 2013 Nissan Leaf does not mean 12 kWh left for you to use — it means 50% of the 21.3 kWh usable portion remains. This is at most ~10.65 kWh (at least when the battery is new.)
3. Operating temperature has a significant influence on battery performance — charging, storage, and discharging — as well as on the expected power usage for climate control units — particularly when it compels the use of their resistive heaters. Since combustion engines can use waste heat to warm their cabins in low temperature operating conditions, this represents a significant advantage for ICE cars, and introduces a source of serious variability between EPA estimates of one propulsion system’s economy versus the other. (Think of it as achieving greater than the aforementioned 30% efficiency, because in cold conditions, not all of the ‘waste heat’ is actually wasted.) Where manufacturers use active thermal management of the battery, there is an ongoing power cost as well — though this loss of fuel economy is intended to make the batteries themselves last longer, so it does offset the impacts of item #4 to a certain degree.
4. Incremental battery degradation over time makes it difficult to compare fuel economy on a percentage-to-distance basis — because, for example, a vehicle which took 90% of its charge to travel 90 miles can very easily, say three or four years later, manage just 85 miles in the exact same conditions and using the same percentage of a full charge. This does not necessarily mean the car has experienced a corresponding loss of fuel economy (though that could be a factor) it may simply reflect changes in the battery’s usable capacity due to expected degradation — i.e. 90% of full storage may not yield the same number of kWh in the latter case.
Most importantly of all:
5. The EPA measures effective fuel economy on a dynamometer in a laboratory equivalent of level ground, meaning the full positive influence of regenerative braking on EV fuel economy is not adequately reflected by the EPA’s testing. Over the course of their lives, both EVs and ICE vehicles obviously experience a net change in elevation very nearly zero — but the absence of a correction for variability in elevation hurts the projected economy of electrics, because elevation variability greatly favors electrics over gas cars. In other words, the EPA views zero net elevation change as a zero-sum game, despite that between propulsion systems it’s not.
6. Idling and stop-and-go traffic jams require a much smaller amount of energy for electric vehicles (also for many hybrids) versus gas-powered vehicles, and this is not fully reflected in EPA testing. In fact, the EPA tests barely take traffic jams into account at all — a fact which is obvious from direct observation of the speed curves in the testing procedures [Note: the EPA seems to have attempted to make some adjustment for this, as the changes to the testing procedures made in 2017 slightly drop fuel economy for some vehicles.] Worse, because the procedures do not appropriately reflect such extremely LOW speed driving, they correspondingly fail to reflect HIGH speed driving as well! The overall average speed of EPA tests (27.74 mph) does roughly coincide (-6.3%) with posted figures on American travel distances and times (which predict about 29.6 mph as the average American driving speed), but the variability in speeds is poorly reflected (e.g. the top speed of EPA tests is 80 mph, a speed which is reached at a transient point during the short-duration very high speed testing component of the model — which itself averages just 48.3 mph.)
The truth is that Americans drive both much faster and much slower than the EPA tests reflect. Quite frequently, in fact. The absence of the leading and trailing edges of the speed curve in EPA models negatively impacts EV economy ratings in relation to combustion car performance.
Worse still, this provides a progressively greater advantage for vehicles the LESS fuel efficient they are to begin with— because to achieve the posted/modeled average speeds despite being locked in traffic jams more than the EPA models reflect requires correspondingly more time and speed on the high end of the spectrum. Because air resistance increases by the square of air speed around the vehicle, and very slow stop-and-go traffic reduces fuel efficiency through wasteful idling, both extremes dramatically decrease fuel economy from the middle range estimates provided by the EPA. Ironically, where this is most evident is in Washington, D.C.:
“Washington, D.C., has the worst gridlock in the country, with commuters wasting 82 hours a year stuck in traffic, nearly twice the national average.” 
In sum, you don’t “make up” for the poor economy of very low speed ICE driving by excessive speeding — you exacerbate the problem. Electrics experience less of an issue on both fronts: they are particularly efficient at low speeds and tend to be among the most aerodynamic vehicles — which helps their per-unit fuel efficiency on the high end.
7. Comparing the latent energy in gasoline to the energy of electricity at a fixed point — the fueling station — discounts the costs of producing and delivering those fuels to a fill up point. This factor represents a massive efficiency savings when it comes to using electricity versus liquid fuel for vehicles.
8. Potential losses “from meter to motor” (i.e. through the charging station and the battery to the point of the motor) NEGATIVELY impact true electric vehicle efficiency—if you make the mistake of thinking you only paid for the “30 kWh” in your battery. Phrased differently, they skew EV dashboard readouts to overestimate your mileage per effective kWh purchased— i.e. you pay somewhat more than you think you do (unless you’re reading fuel consumption at the meter.) Conversely, if you’re using the dashboard readout to determine how much fuel you’re using and to project how much driving will cost you, you’ll find you’re paying significantly more at the charger — and even this number varies with the type of charger you’re using!
Items 1–4 above make it difficult for users to know what their precise fuel economy is or should be, while items 5, 6, and 7 are the three main reasons why the EPA’s “MPGe” ratings significantly discount electric vehicle fuel economy versus user reports (8 adjusts the figures back in the direction of the pessimistic EPA ratings.) To see how electric vehicles are unfairly weighed by the EPA first requires looking at the test standards and at gas powered vehicles rather than electrics…
What you CAN’T see here is any sort of model to reflect the impact of changes in elevation on a vehicle’s performance— a factor which greatly favors electric vehicles over ICE vehicles. Changes in elevation are simply not built into EPA laboratory tests, despite that they influence essentially all vehicles in the real world (outside of a track or laboratory.) Yes, all vehicles require more energy to go up hills, but only vehicles with regenerative braking can consistently recoup a significant percentage of this energy when returning from elevation — and whether you realize it or not this is constantly happening to some degree, no matter where you live. Essentially, test dynamometers are excellent at applying variable loads to vehicles in order to simulate forces acting on a vehicle in motion, but they are poor at measuring the energy electric vehicles recover *whenever* they are returning from elevation.
Roughly speaking, that’s about half the time.
While it’s undoubtedly true that EPA models use dynamometers capable of simulating regenerative braking for stop-and-go conditions (I could not conclusively verify this) the conspicuous lack of modeling for elevation changes for gas cars implies a corresponding absence in the tests for electrics, which therefore implies exclusion of down-slope modeling. (Incidentally, if the tests for electrics actually DO have such a correction, it begs the question of why gas-powered cars clearly don’t have it — which very obviously leads to a drastic loss of real world efficiency versus EPA test estimates in that going up and down hills definitely costs more fuel.) In short, this represents a significant disadvantage for EVs when it comes to EPA testing versus reality.
The key takeaway here is that despite prevailing myth, electric vehicles become progressively more efficient than combustion vehicles as elevation variability increases. Especially when those changes are significant, and regardless of whether they occur during a trip which ends in a net elevation change of zero! The main reason for this has been noted: it’s because while all vehicles require additional energy to climb, electrics waste far less energy on controlling their descent. It just isn’t a simple matter of the speed the vehicle is traveling — it’s where the energy for traveling that speed comes from, and potentially where it goes. Anyone who has crested a steep hill at or near the speed limit should find this very obvious, but just in case…
Downward slopes in excess of about 5-7% tend to accelerate vehicles beyond tire and aerodynamic resistance at typical highway driving speeds without their drivers having to apply fuel. Such descents require intermittent application of the brakes for combustion vehicles (with corresponding energy losses) while electric vehicles can convert much of this energy back into usable form by charging their batteries. While this impact is most pronounced at high grades, it occurs on lower grades as well. This is the first of two reasons why EVs are more efficient in mountainous areas rather than less efficient. The second reason involves changes in effective fuel-air mixture — which can potentially reduce motor horsepower for combustion engines at elevation, but which do not impact electric motors similarly. Simply put, EVs are not bothered by thinning air, and as such can potentially gain a comparative advantage at elevation. This point is worth expounding on to some degree…
EPA models, which are tested in Ann Arbor, Michigan at 840' versus a U.S. national average elevation of 2500'  assume functional oxygen sensors — the devices which monitor exhaust gases to determine whether the proper fuel-air mixture is being utilized . These devices, in conjunction with a modern combustion car’s computer, help to minimize pollution and ensure maximal fuel efficiency. Unfortunately, as shown here , oxygen sensors frequently rank at the top of required repairs — taking the #1 spot in reference year 2015 at an average cost of $249— and are generally NOT a repair which is necessary to keep a vehicle functioning (it is a repair which is necessary to pass a smog check, however.) What this means is that whenever a given combustion vehicle is not operating under the specific barometric pressure its computer is set to AND has a faulty oxygen sensor — a situation which is remarkably common — it will under perform in comparison to an electric vehicle until it is fixed. The dashboard check engine light will probably come on, but a driver is free to continue operating the vehicle at a lower efficiency — and may not even notice the reduced efficiency (to say nothing of the added pollution.) Furthermore, the average cost of replacing an oxygen sensor is far higher than the value of the fuel a driver would waste in a full year — even if you assume a 5% loss in fuel, $4/gallon gasoline, and a vehicle whose normal fuel economy is only 20 mpg. This economic forcing function encourages people to forgo the cost of replacing faulty sensors at least until the next scheduled smog check, and in at least twenty states, there are no compulsory smog checks to compel the replacement! (Most states have annual checks, though some have biannual checks.)
EV motors are not influenced by this factor at all, while taken together, ICE vehicles will NEVER trend to the average efficiencies predicted by the EPA tests over their lifetimes — because to do so would require either operation at constant elevation, oxygen sensors which uniformly last for the life of the vehicles, or both. This is one of several reasons (misfiring spark plugs is another) why using the EPA estimates to predict lifetime fuel costs can seriously underestimate the true costs, and skew the analysis in favor of combustion cars. It would be fine to disregard this fuel consumption reducing factor if it impacted both electrics and gas cars in the same way, but it doesn’t.
As a side note, hybrid vehicles enjoy similar advantages in regenerative braking (and often in idling.) This is why the standard mpg ratings for hybrids similarly don’t reflect their actual performance in relation to fully gas-powered cars—i.e. the hybrids, while still suffering from reduced combustion efficiency over time, are still marginally better than face-value comparisons of EPA numbers suggest. (Note: plug-in hybrid MPGe estimates, however, are extremely misleading. They only apply when the vehicles are expressly using electricity, and the true efficiency you get from even that electricity depends on where it came from — a charger or the regenerative braking [at least if you only want to look so far back as a single step in the power production chain.] It’s quite literally impossible to track where the electricity in the battery came from in the case of a PHEV, so you’re basically left with a car that gets noticeably better gas mileage than a normal combustion vehicle, but acts as its own terribly inefficient power plant whenever its user decides not to plug it in. In that case, you sure can’t honestly compare its MPGe efficiency to an electric car, because in the worst case an electric gets its power from a far-off coal plant somewhere, while the PHEV is tooting along on electricity it made for itself with a tiny gas motor running at high rpm which its owner fed into the car’s kinetic energy pool and then pulled back out using less-than-100% efficient regenerative brakes. I don’t even want to know how inefficient that is, but it sure isn’t anything close to 50%. Come to think of it, in that case MPGe will be completely irrelevant, because we’re already talking about a car you don’t bother plugging in — so the only thing that will matter is your gas pump fuel economy. This is why the best you’re going to do with a PHEV is a weighted average economy between the MPGe and its gas-only rating, depending on where you got the fuel from.)
Moving back to strict EVs versus gas cars…
The refining cost —in terms of units of electricity used — to produce one gallon of gasoline has been computed here  as approximately 4.5 kWh. This figure must be added to the effective energy of a gallon of gasoline to fairly compare electricity values to it and compute an “MPGe.” This, then, makes a true equivalence. This calculation doesn’t diminish the value of a gallon of gasoline once it arrives in a vehicle’s tank, but to suggest that such a gallon’s 33.7 kWh can be accessed without first investing 4.5 kWh in refining it is just as untrue as believing an electric vehicle can hold 40 kWh of charge without drawing more than 40 kWh from an outlet in order to do it.
Yes, a gallon of gasoline can enable a Prius to travel 50 miles, but that gas cannot even be refined without using 4.5 kW of electricity — thus the effective energy of the fuel is lower than its chemical bond energy because it is carrying production debt. Viewed in reverse, we could as easily describe the distance an ICE car travels based on the supposed chemical bond energy in gasoline (33.7 kWh) — for example to say a 50-mpg Prius makes 1.48 mi/kWh — and then compare that to the economy experienced per kWh from the battery of an electric (4.1 mi/kWh in the case of my 2013 Leaf.) Doing so, we would find an effective 138.5 MPGe versus the EPA’s estimated 115 [because (4.1/1.48)*50 = 138.5.]
The comparison between economies must begin at the level of the vehicle, for obvious reasons. To fail to do so is to give the general public an erroneous impression of the efficiency of the cars, and of the impact of using them in proportion to the impact of using combustion vehicles. Thus to determine how economical a vehicle is, take the effective net energy it is holding and figure out how far it can travel with it. For an EV, this is the only reasonable approach — because outside of the reading on your instrument panel, it’s generally impossible to know how much fuel you’ve dispensed from your home charger into your car. Well okay, Tony Williams, it’s certainly nowhere near impossible but let’s face it, who is going to install a charger which is more expensive (because it has an inline dedicated electrical meter) when it’s just as easy to simply take the reading off the dash and multiply? [I’m kidding, Tony, I knew you knew that.]
The MPGe rating, then, must begin at an effective 38.2 kWh per gallon equivalent rather than 33.7 kWh. This establishes a level playing field for the comparison, and effectively results in significantly higher — and fairer — MPGe ratings for electrics to this point in the analysis. If we assume that the delivery cost (source to customer) per unit power for gasoline versus electricity is equivalent for the two fuels (an assumption which still puts electric vehicles at a massive disadvantage, since it costs far less to transport electricity than it does to transport any sort of fossil fuel — the latter of which is nearly impossible to quantify), the only remaining factors to adjust for are, from :
- The efficiency of the charger (95%)
- The efficiency of the inverter (95%)
- The losses inherent in pushing electricity to the battery — the largest of which is heat (90%)
We will not deduct the efficiency of the electric motor/drivetrain, because the above factors will be used to adjust the EPA stated economy in relation to real-world experience to result in estimated at the outlet costs. What this basically means is that when you have, let’s say, 42.4 kWh of usable power in your Chevy Bolt battery, you’ll actually have paid for a total of 52.2 kWh (42.4/0.8122) at the meter. If this sounds discouraging, wait for a moment until you see what the final costs of using gasoline versus electricity are.
The most remarkable aspect of this table is that it shows EPA estimated “MPGe” to be significantly different from (and lower than) the true values. The true values, in green at the far right, demonstrate an apples-to-apples comparison of electric vehicle performance versus combustion cars. That’s right, the AWD Tesla Model 3 is roughly three times as efficient as a 50-mpg Toyota Prius.
Since the MPGe rating is primarily used to inform the public about the relative efficiency of electrics versus combustion cars, the EPA’s usage up to now has been grossly misleading. Many members of the public want to know how much better they’re doing by the environment, and inappropriate labeling makes it difficult for them to do this.
The other aspect of this chart which bears repeating is that the values in the third column from the right — “Corrected from wall” — reflect the number of miles you can expect per kWh of electricity you actually use for your car, and the number under “From battery” is the mileage you can expect per usable kWh once it’s already been delivered to your vehicle’s battery [i.e. the number you normally reference.] No, all of this analysis didn’t change the fuel efficiency much between what you’ll see at the meter and the derived mi/kWh EPA ratings from the ubiquitous Monroney labels at dealerships, but the EPA’s MPGe ratings are 10.6% lower than they should be, and people use those figures in part to make purchasing decisions. Perhaps worse, people typically compare the EPA figures to their dash, not to their meter, and range is calculated not from metered power use (EPA) but from the battery (users.) At least now you can estimate the net impact of your mileage on your electrical bills should you choose an electric car.
As a bonus, now you can also compute realistic range:
For a new 24 kWh 2013 Leaf S (21.3 kWh usable), 21.3 kWh * 4.1 mi/kWh = 87.3 miles of range, average, and would require 21.3/0.8122 = 26.2 kWh at the outlet. This would cost $3.41 at the current national residential average of $0.1301/kWh, or the equivalent of 1.354 gallons of gas at today’s (2/26/18) national average price of $2.517. The 50-mpg Prius would cost almost a dollar more, at $4.39, to travel that distance, and the 2018 Honda Accord in our test? $8.45.
Did you know you were paying a dollar more every 88 miles with your Prius, and FIVE dollars more every 87 miles with your Accord?
Full battery cycles with the ’13 Leaf starting and ending at the same geographic point seem to concur with this 87-mile-range estimate, varying from an extreme low of about 65 miles in the dead of winter (with a new battery) to a high of over 120 miles during a 70 degree summer day. They settled on an average around 85–90 miles, as reflected by the long-term fuel economy indicator, which read 4.1 miles/kWh.
The critical takeaway here is that while the derived EPA numbers are very close to what you can expect if you compare changes in your electrical meter to the distances you travel, in practice drivers rarely (if ever) read the meter on the side of their house when they’re trying to figure out how far their car will go: They look at their dashboard. This is one reason why users consistently report effective range longer than posted estimates — the EPA effectively measures fuel consumption at the wall, whereas the distance an EV can travel is based on how much is stored in the usable portion of the battery.
Lastly, with an electric vehicle (or hybrid), you’re provided an extremely helpful additional tool for enhancing your fuel economy: the ‘mobile charger’ of your regenerative brakes. Even without changing your driving style, you become a more efficient driver because you waste less kinetic energy in slowing/stopping your vehicle. When you master the skill of stopping without appreciably using the dedicated brakes, you’ll achieve even better economies.
This next table shows what the electrics would cost per year if they required no maintenance to reach ~200K miles (202,110.)
We’ll get to those maintenance requirements in a minute, but before doing that, let’s see how the cars stack up side by side — by ranking them all, solely with MSRP and fuel consumption taken into account (year 15):
As you can see, when you include rebates, the electrics dominate the list — taking 14 of the top 15 spots. Several of the electrics stack up nicely to the most affordable gas car of the bunch (the Civic) — and four of them beat all of the other combustion cars in the list even without considering rebates.
EVs seem to have already reached cost parity — at least if you consider long-term ownership costs. What does “cost parity” mean, if it doesn’t mean long-term ownership costs?
So with full rebates, all of the base electrics tested beat all of the gas cars — at least prior to accounting for maintenance. We’ll get to how those rebates work and how many people might still be able to claim them toward the bottom. For now let’s turn to maintenance, the final big X factor…
It is obviously extremely difficult to quantify what the total maintenance costs of vehicles will be. However, we can make general estimates of gas car maintenance using the reference here  — which indicates:
“ A stable, consistent increase of $150 per year in costs exists for years 1 through 10. After that, there is a distinct jump between 11 and 12 years of age. After age 13, costs plateau around $2,000 per year. This is likely because people disown their cars if maintenance costs are higher than their cars’ worth.”
The same site also lists the Camry, Rogue, Civic, and Accord as costing $5,200, $6,500, $6,600, and $6,600 [by year 10] respectively as the least expensive of the combustion cars in this comparison (and among the least expensive cars to maintain, period.) The remaining cars from the test list above will have their repair costs set to $7,500 at year 10, as this is the boundary lower than which they would have appeared on the same “least expensive” list in  — thus it represents the most competitive estimate allowable from the reference cited. Annual maintenance costs average over $1500 per year between year 10 and year 15 according to this diagram, referenced on the same site:
I’ve used the comparatively less expensive repairs of the Camry, Rogue, Civic, and Accord to reflect similarly reduced costs for those vehicles from year 10 to year 15, reasoning that the above table shows average vehicle maintenance cost and that the parts and labor for these models are historically lower. [For each of the vehicles, I effectively adjusted the year 10 through year 15 repair costs to reflect the average rate of repairs in years 0–10.] The year 6 costs reflect the “stable consistent increase” referenced above. The following table presents this:
Now for the tougher question: What will the electrics cost to repair?
Let’s break it down into the main high-ticket items and the items which will regularly need to be replaced — e.g. wipers, cabin filter, & tires — and use the comparatively more expensive Tesla as a reference for the EVs.
- The car itself is covered by a 4-year, 50,000-mile basic warranty — roughly comparable to industry standards, though a bit better than some manufacturers.
- The battery and drivetrain for the Model 3 are covered by an 8-year, 120,000-mile warranty for the long range version, and an 8-year, 100,000- mile warranty for the short range version. This covers degradation beyond 70%; if the battery degrades below that point, Tesla will replace it free of charge. 
- Low rolling resistance tires wear faster than regular radials and are therefore generally more expensive per mile than standard tires. They are also usually factored into the fuel economy estimates, so if you don’t use them you’ll cost yourself economy and likely experience a net loss of money. Because you can usually expect faster wear on tires for an electric car than for a gas car, I’ll use an estimate of 25% higher tire costs per mile for electrics versus gas cars.
- Cabin filters and wipers are common to both gas and electric cars. They will cancel out as they are not significantly more expensive for one than for the other.
- Brake pads wear significantly slower on electric vehicles with proper use — because regenerative braking recovers much of the car’s kinetic energy back into the battery without it being wasted in friction to the disc brakes. We’ll cut the total lifetime brake job costs by one-third for the electrics versus the gas cars (this factor varies considerably depending on driving style.)
- There are no exhaust systems, no fuel or water pumps, no spark plugs to replace, and the overall number of moving parts in an electric vehicle — those most prone to wear and breakage — is at least 10–20X smaller for an electric than for a gas-powered car. The table below shows the components from this list which are associated with routine maintenance, and how much of a role they are likely to play in the costs.
- The 12V battery, which supplies power to the electronics, etc. will typically need to be replaced more often in an electric vehicle. We’ll consider change intervals of 30K miles versus about 40K miles for this component. 
- The optional extended Tesla warranty on the Model S/X is a good reference point from which to determine what the ongoing costs of repair might be from the end of year 4 through the end of year 8 for the Model 3. (Tesla does not currently offer an extended warranty on the Model 3, however. I’ll cover why they absolutely should in another post.) For this reason, I’ll estimate the year 4 through year 8 costs as $3000 —two-thirds of the midpoint of the extended warranty for the Model S and Model X — which seems a pessimistic estimate for a car which costs about half as much as a typical S or X. Additionally, extended warranties act as repair insurance for vehicle owners and are generally priced by manufacturers at or above the expected repair costs a user will on average experience — which effectively means estimating true costs by using existing S/X warranties again seems somewhat conservative.
- After year 8 passes, the cars in this example will be at 107,792 miles and out of warranty. For the moment, I have ignored the fact that vehicles in this test will reach 100,000 miles (end of warranty) roughly 40–45% of the way through year 7. [This amounts to a total difference of $300–470 depending on the model, according to the calculations below.] I will estimate a yearly average 10% chance of total battery failure for the electrics, which amounts to an aggregate 52.2% chance of the battery failing during the non-warranted period between the 8th and 15th years. From this figure, I’ll compute expected value (cost) of replacement in any of those given years based on the projected costs of lithium ion batteries in those same years.
- Tesla has offered an eight year unlimited mile warranty on the drive train in the S/X. For reasons detailed below, I’ll estimate a constant 5% annual chance of total drivetrain failure per year from the end of year 8 to the end of year 15 (which amounts to an aggregate 30.2% chance of drive unit failure for the single motor vehicles, and — assuming independent chances of failure — a 60.4% chance of one or the other of the AWD Tesla’s electric motors failing in that same period. Elon Musk would no doubt be irked by such a pessimistic estimate, but until we see data, or until Tesla offers an extended warranty plan which can indirectly tell us what sort of reliability the company itself expects out of the Model 3, we can’t reasonably be expected to accept promises of ‘million mile drive trains.’ Never let it be said that I didn’t offer an impressively high bar for EV manufacturers to vault.)
The costs in 9 & 10 above will make up the bulk of the EV maintenance costs between years 8 and 15, and when added to the distributed lifetime costs in the leftmost column of table below will represent the estimated total maintenance costs for EVs.
Okay, to present this in a more visually obvious form:
The table above shows the primary lifetime repairs-in-common, and is meant as a rough starting point from which fears of EV drivetrain/battery failure can rationally be discussed. The column for ICE vehicles shows a common regular maintenance cost of $10,500 —almost exactly the total projected year 15 cost of maintenance for the least expensive ICE vehicle in this comparison: the Toyota Camry’s $10,400. This obviously doesn’t include many of the most common repairs, and each of the repairs listed can clearly vary widely in price from one vehicle to the next. The table below shows the most common vehicle repairs in 2015 — the majority of which don’t show up on the list above:
- Replacing an oxygen sensor — $249
- Replacing a catalytic converter — $1,153
- Replacing ignition coil(s) and spark plug(s) — $390
- Tightening or replacing a fuel cap — $15
- Thermostat replacement — $210
- Replacing ignition coil(s) — $236
- Mass air flow sensor replacement — $382
- Replacing spark plug wire(s) and spark plug(s) — $331
- Replacing evaporative emissions (EVAP) purge control valve — $168
- Replacing evaporate emissions (EVAP) purging solenoid — $184
To be clear, the range in total maintenance costs for the life of the vehicles in this analysis has been estimated from a low of $10,400 for the Camry to a high of $15,000 (shared by six of the vehicles) elsewhere. I’ve merely presented the numbers in a different format, and estimated how those costs mount over the last four years of the vehicles’ lives. The combustion car maintenance estimates in Figure 10 above are merely intended to show how the routine maintenance costs typically add up over the life of the vehicles without even considering the most common repairs. I’ve equated the routine EV repairs line by line to show that I haven’t introduced bias here. I’ll be adding critical repair costs to the EV side of the list, and estimating additional costs for the combustion cars as between about $0 for the Camry (which saves enough in affordable routine costs to make $10500 last a full 15 years), to $2500 (for the Rogue), $2700 for the Civic and the Accord, and a maximum of $4500 for the other six. This number must include eight of the ten most common vehicle repairs beyond routine maintenance for the life of the vehicles, and account for any chance of catastrophic failure (engine, transmission, etc.) as well.
*Importantly, the chance of catastrophic engine failure for a combustion engine is significantly higher than the chance of failure of an electric motor. In other words, the dollar costs of the repairs in the second clause of the last sentence of the preceding paragraph are likely to be considerable.* Sure, motors can be rebuilt — but imagining that electric car repairs cost some absurdly large multiple of ICE car repairs because you don’t even know what’s in them is an example of argumentum ad ignorantiam. You can’t just assume something is true simply because it hasn’t been proven false.
As Ben Sullins likes to say, free the data, your mind will follow.
The final piece of the vehicle data puzzle comes in the EV battery and drive train values themselves, and their potential for failure outside the warranty period. This is not a straightforward comparison, because battery costs have been falling and are projected to continue to do so due in large part to efficiencies of scale and improved manufacturing techniques — advantages which the ICE vehicle industry has already implemented for their components to a large degree. [Solid state batteries could drop production costs even further and/or significantly improve cycle life — and may well be available within the next fifteen years.]
2018 Chevy Bolt
Currently, GM lists the battery replacement cost for the 2018 Chevy Bolt at a whopping $15,734.29 — for a 60 kWh pack  — this represents a rated-unit capacity cost of $262.24/kWh. This number appears out of scale with the other manufacturers, considering the competitive numbers it has released on battery purchases from LG chem ($145/kWh). I’ll use this figure anyway, and leave the reader to speculate as to why GM’s mark up and battery shell are so high. It’s certainly hard to guess how GM arrives at the figure, given that EV manufacturers generally require battery exchange during the replacement process, meaning that aside from the exchange value difference in batteries themselves, the primary cost factor in replacing a battery will generally be the labor. [For reference, Tesla was offering 85 kWh battery replacements about five years ago at $12,000 — a unit capacity cost of just $141.18. This is no doubt again related to the requirement of a battery exchange, and the fact that Tesla produces and can recycle its own batteries.] We’ll return to the Bolt in a moment, for now let’s look at the Tesla and the Leaf.
TESLA Model 3
According to this source , Tesla’s per-kWh production cost of batteries was about $124 approximately one year ago. That figure yields a production cost (cells only) of:
55 kWh (short range battery): $6820
75 kWh (long range battery): $9300
Assuming a 25% mark up on both of these figures from internal production to delivery yields a cost new of about $8500 for the short range and $11,600 for the long range battery. In light of previous 85 kWh battery replacement costs ($12,000) and the fact that replacement requires exchange of the used battery — which retains a significant portion of its value (particularly for a battery manufacturer) — these figures seem in line with expectations. For this discussion, I’ll estimate a current replacement cost of the long range battery at $11,600 and $8500 for the short range battery (Tesla has not yet posted figures) and include the exchange of the used battery.
2018 Nissan Leaf SV
The replacement battery prior to the 2018 model year cost a straightforward $5499 plus an estimated three hours’ worth of labor  and forfeit of the vehicle’s old pack. Assuming a $100/hour cost of labor, that amounts to $5799 for the 30 kWh battery the car previously had, and implies a replacement cost of about $7650 for the 40 kWh in its current model. Checking this figure in relation to the Model 3’s numbers above yields a cost of $10400 if Nissan were using a 55 kWh battery, and $14050 if it were using a 75 kWh battery — numbers about 20% higher than the projected Tesla costs for those capacities—reasonable for a company which last fall sold its battery production division, is no longer in the business of making their own, and had more expensive battery production costs to begin with.
[From these figures, it appears as though something funny is going on with GM’s battery production: the replacement cost seems far too high. The simplest explanation I can come up with is that it hopes to profit on battery replacements, since it does not currently profit on the car itself (setting aside ZEV credits.) It may also be having manufacturing difficulties of some sort. Let’s hope its battery replacement costs become more competitive with the other two market leaders — since its battery warranty is the least robust of the three. This seriously impacts the long term costs of operating a Bolt, as shown below.]
According to this article  lithium-ion battery production costs have dropped some 80% in the past 8 years according to Bloomberg and are expected by most manufacturers to drop below $100/kWh by 2025. I’ve used more conservative estimates in the table below — which shows the current cost by manufacturer along the top row (Leaf numbers are estimated based on published per kWh values and prior battery replacement costs) and thereafter the projected costs in years 8–15:
The logic behind the above table may not be immediately apparent. Why, when the average cost of replacing an EV battery in the referenced years ranges from an average of about $5,450 (for the Leaf) to $11,500 (for the Bolt) is the replacement cost listed as far lower ($2,879 to 6,076)? The reason is that the comparison is based on a 52% aggregate likelihood of this repair being required over the specified time period. This is essentially the statistical adjustment necessary to reflect expected values. If this weren’t done, it would effectively mean that you’re assuming the battery will need to be replaced in all cases, which is obviously an erroneous assumption.
In actuality, the odds are certain to be lower than 52% if the experiences of Model S drivers are any indication. Battery exchanges have been exceptionally rare even for drivers who have already exceeded 150,000 miles — a mileage at which the model we’re using estimates an approximate 30–32% chance of necessary replacement. Of those drivers who have bested 150,000 miles to date, far fewer than one in three have needed replacement of their batteries.
The last piece of the puzzle — the cost of the drive motor— may seem to be a great fearful unknown for those who have never considered an electric vehicle. What happens if my drive unit fails?
For this, we’ll begin with an estimated drive motor replacement cost of $5000. The reason for this is simple: the AWD Tesla Model 3 has been estimated at being $4000 more expensive than the RWD Tesla Model 3 — i.e. a second drive motor essentially costs an additional $4000. Factoring an additional 25% for miscellaneous associated issues and labor costs yields the $5000 we’ve estimated for total motor failure. Less easy to predict are the chances of total motor failure occurring. I could find no data on either gas or electric vehicle motor failure rates from which to base my estimate. Ultimately, I settled on an annual chance of 5% — half that of battery failure above — which as previously mentioned results in an aggregate chance of failure for the single motor electrics of 30.2% over the 8 to 15 year period tested, and effectively a 60.4% chance of one or the other motors requiring replacement for the AWD Tesla. I also factored in a slight correction for the case of both motors failing at the same time for the AWD Tesla, as you can see just below in the next table.
We reach a hang when speculating about the cost of a Chevy Bolt or Nissan Leaf motor, but we do have figures for horsepower — and we know that motor horsepower has a greater-than-linear relationship to motor cost. We’ll use a linear estimate, noting that this does result in a slight disadvantage in the comparison for the Bolt and Leaf versus the Tesla.
This information, taken together, yields the following chart:
We can now calculate the expected total cost of the vehicles in this comparison in any given year. Here is the full chart, I will follow it sorting the information year-by-year:
Shocking or not, this table reflects realistic total costs of ownership at year 15 — costs which could drop even more for the electric vehicles through the use of home solar. The reality is that despite being not as familiar to the general public (and having less familiar maintenance costs) electric vehicles on average cost slightly less to maintain than combustion cars. This further extends their well-known full cost savings. According to the results here, if you consider lifetime costs, electric vehicles are already at cost parity with their gas-fired counterparts — even before rebates are considered.
Figure 14 shows the rankings after just six years, without rebates, at which point the least expensive electric is 41.5% more expensive per mile than the Civic, but still less expensive than 60% of the gas cars in the lineup.
Figure 15 moves on to the end of year 11, again without rebates. At this point, the electrics make up five of the top ten slots, and are catching up in per-mile operating costs (rightmost columns.) This reflects the front-loaded costs of operation for electrics , and how they are recovered substantially over time.
In Figure 16, below: after 15 years, no rebates — the Civic edges out the Leaf by just under 1% — less than $22/year.
So after all that, do the gas cars basically win? Seems kind of anticlimactic…
True, if this were the Olympics, the Civic would still walk away with a gold medal. Plus, with gas-powered cars there is that sense of familiarity, and you don’t need to worry about battery degradation or range…
Unfortunately, using familiarity as the ultimate Trump card will cost you greenbacks (even before you consider any impact on the environment.) In the above comparison, I’ve been more than fair in the cost analysis across the board — in fact I’m sure I’ll receive hate mail from electric vehicle advocates (which is a bit of an oxymoron, actually) for weighing things far too heavily in favor of combustion vehicles. In reality, the long term costs undoubtedly favor electrics by a comfortable margin, for the following reasons:
- Efficiencies of scale for EV production have not been achieved yet — they’ve hardly even been approached.
With the exception of Tesla Motors, which is constantly being derided in the press for “burning cash” (=investing money in infrastructure to build the cars of the future), there are few (if any) manufacturers in the U.S. prepared to produce a significant number of electric vehicles for the next several years (Nissan and GM might be exceptions, but both are still seriously wrangling with the question of producing an EV at a per unit profit — they cannibalize profits from combustion motor sales to fuel EV R&D and production. This likely means they’ll be reluctant to step up production at least until per-unit breakeven can be achieved. This is what we currently see happening. It’s the reason why we see virtually no marketing for any of the electrics — because like it or not, there are TWO significant determinants of how much money it’s worth spending on marketing, not just one. People, and profits. It’s the number of people who might be interested in the vehicles TIMES the net profit per unit — hence when you’re losing money on every car you sure don’t want to spend even more money trying to encourage more people to buy it!)
Anyone with a fourth grade education knows that a positive times a negative is a negative — and adding another negative to the equation doesn’t help. Cute trick there, Ms. Barra, but people are watching you.
While critics are quick to point out that “Tesla hardly makes any cars” [and from the standpoint of total market share, that’s true] what they tend to leave out is that the company is within reach of producing ten times as many electrics as any other domestic automaker currently builds. They will continue to be quite willing to produce more, because they’re actually profiting on unit sales — it’s just not as visible because up to now they’ve had to use those profits to build essentially all of their manufacturing infrastructure. To catch up to companies that have been chugging essentially the same exact vehicles off their production lines for the last 40+ years (albeit with a few cosmetic changes here and there.) Efficiencies of scale in this market have yet to be approached, and real competition has yet to enter the fray. As a result, the costs of EVs will continue to fall.
2. Grid-scale photovoltaic energy production is trending below $0.05/kWh lifetime levelized cost, making it the least expensive form of energy production to deploy.
It is easy to think that prices never fall — that they always go up — but with technology refinements this isn’t always the case. The cell phone is probably the most familiar example (costing almost $4000 when it debuted  and dropping to about 15% of that price for a modern version — which has computing power many orders of magnitude higher), but falling PV prices (a curve described by Swanson’s Law ) are an example as well. This could ultimately mean overall electricity prices which actually drop over time as photovoltaic production emerges as the primary source of electricity (as of 2016, the fraction of electricity from solar was a meager 0.9% .) At the very least, in terms of price based on reference year, the gap in price of fossil fuels versus photovoltaic energy will certainly continue to widen. The former are inevitably getting more expensive and the latter continues to fall in price.
3. The residual value of an electric car is significantly higher than the residual value of a combustion car, when all factors are considered.
This is probably exactly the opposite of what you’ve heard, but hear me out for a minute (okay, for another minute.) What you probably already know is that the U.S. market values of used electrics are — by percentage — significantly lower year-by-year than their gas-fired counterparts. That is, an electric’s value after year 1, 2, 3, etc. is significantly lower as a percentage of its starting MSRP than a comparable gas-fired car. In part, the reason for this is that the vast majority of the electric vehicles in current circulation in the U.S. took advantage of the Federal tax incentive of $7500, dropping their effective MSRP accordingly. When a user sells such a car, this must be accounted for in the sale — not necessarily because he or she effectively realizes that the original price paid was already up to 20% less than the MSRP, but because prospective buyers are faced with the choice of buying a used EV (without a rebate) or a new one (with a rebate.) When you compare the residual market value after correcting for the upfront influence of the rebate, the difference between the market value for a used electric car and the corresponding market value for a used ICE car is much smaller.
Worse, because this factor is not widely understood in the general public, a comparison shopper will see the lower market price of a used electric in comparison to a gas car which initially carried a similar MSRP and somewhat paradoxically view it as less valuable as a result! This principle is covered very well in this video:
Another hard-to-notice influence on market value is the “love them or hate them” nature of electric vehicles. The majority of people who buy an electric vehicle would do so again…90% of them would according to research by Ford cited here: . However, reduced or inadequate range (and manufacturing defects on some of those vehicles) makes a certain percentage of them unusable over time to their initial buyers. Somewhat paradoxically, this results in a preferentially lower market value for used electrics — because as a percentage, fewer of the initial buyers are willing to give up their cars, and a correspondingly greater percentage of the cars which do hit the market are given up because they have manufacturing defects, or insufficient range for example. This market value — even after being adjusted for rebates — will then be effectively lower than for comparable gas cars, which tends to result in an erroneous view of the residual value of an electric vehicle. Used market value <> residual value, nor are the two perfectly correlated (i.e. halving one won’t necessarily halve the other.) Residual value — that is, the value of a car to its owner — is the value it still has in relation to how much its user originally bought it for, which is quite different than its market value. Even though we use the terms interchangeably, they’re not the same thing. This is basically the trash/treasure principle, and it has a big impact on used market values for electrics in part because a very large fraction of people don’t buy electrics solely based on their price. They buy them because of other factors which can’t be obtained from petroleum-based cars, like quieter operation, smoother acceleration, zero tailpipe emissions, and the ability to fuel their car from the roof of their house.
Think about it: if market value were exactly equal to residual value, it would effectively mean that any given owner sits precisely on the fence with respect to keeping his or her car or selling it (discounting transactional costs.) In reality, the vast majority of people value their cars much higher than the market value — and in electric vehicles, this impact is even more pronounced (a fact which the Ford study suggests, but does not explicitly find.) The other reason why it must be true that the residual value of owned electrics is disproportionately greater than non-electrics is that their costs of operation are heavily front-loaded — and the transactional costs of buying one (many of which are determined by percentage multiplication of those front-loaded costs) are therefore higher as well. This financially discourages owners from selling their vehicles until those higher non-refundable costs can be recouped through operation. In effect, users don’t just hold onto their electric cars because they love them — they hold onto them because they know doing so will invariably continue to save them money, which is itself a component of their residual value!
It bears mentioning that market value is significantly depressed for cars which are bought and sold more frequently per unit time — because the more frequently cars of a specific type are bought and sold, the greater the accumulation of transactional losses on the specific model in question, and the more likely it is the model in question has an out-of-the-norm defect or unusually high repair costs, for example. This is a selection effect. What it essentially means is that the main driver of the market value of used vehicles — aside from MSRP, age, and mileage — car for car, has historically been reliability and repair costs.
Now I know you’re probably saying, “DOH! Tell me something I didn’t know.”
Okay, how’s this:
You will very rarely, if ever, find a gas-powered vehicle with a value which is significantly depressed merely because of how far it can travel on a tank of gas.
Yeah, I know. You knew that, too. But to bring it back to the discussion, recall that we were talking about residual value — that is, the value to an owner — and market value — that is, the value of the car if you tried to sell it — and how electric vehicles vary in this regard versus combustion cars.
The big secret?
A significant portion of the value of an electric vehicle depends on how far it can go on a charge, which means that if it hits the used market because it can’t go as far as its user needs it to anymore, it doesn’t mean it doesn’t retain REAL value for you much higher than the value it had to its former owner, it just means that the user couldn’t help but sell it because it couldn’t do what he or she needed it to anymore.
This is trash/treasure again. If I don’t ever drive more than 30 miles per day, for example, I’ll snatch that $5000 2013 Nissan Leaf out of your hands in a heartbeat — even if it has only 8 of its 12 capacity bars left. It won’t matter to me that you needed to drive it 60 miles to work in the dead of a Minnesota winter and your boss won’t let you plug it in. The only thing I care about is the price and whether it will work for me. That’s the primary reason why market value is still a very bad predictor of what the actual value of a used electric car is: because most people are still using the if it’s cheap and used then it’s not a good car philosophy and compounding how much they discount electric vehicles on that market (because they’re electric) rather than evaluating each individually based on whether they’re still more than adequate for their needs (most specifically their usual driving distances.)
On the one hand, this is understandable. Used market value has historically been guided in large part by reliability and repair costs (in addition to things like the age of the car, mileage, brand, etc.) — and unusually low values for a particular type of car in a particular model year have generally meant a less reliable or more costly to maintain car. Adding the two additional factors mentioned above (the depression of cost due to the Federal rebate and the possibility that an electric is sold by its original owner because it didn’t have sufficient range) depresses the market value of used electrics and makes them appear to be worth quite a bit less to a population which is accustomed to judging value based on price.
But while cars which have low market prices tend to be viewed as cheaper (=less reliable) in general, this doesn’t impact the value to an owner who isn’t interested in selling one, except tangentially. It’s actually likely to lower insurance costs, which are determined in part based on the car’s current value.
The extremely low prices on used Nissan Leafs in 2015 and 2016 in particular were evidence of this. In effect, the used car prices were much lower secondary to the Federal rebate, people in the U.S. shied away from them as being inferior to gas cars of corresponding model years, and then — beginning in late ’16 (possibly triggered in part by this article: ) foreign buyers noticed that Americans didn’t realize the value of the cars sitting right in front of them and picked the market clean. This is understandable on the grounds that most countries drive, as we do, on the right, most have much higher fuel prices than we do, and most drive half as much or less than we do — meaning range in a used electric is likely to be much less of an issue for foreigners than it is for U.S. consumers.
Our trash became their treasure.
One final point on residual value…
On top of all of this, the scrap value of an electric car is higher at end of life than the scrap value of a comparable gas car, because the scrap value is ultimately determined by the materials it is constructed of, and how much they cost per unit mass. This means that electric vehicles — having batteries with material costs per unit mass higher than the material costs per unit mass of combustion motors — will get you a few more dollars once you’ve driven them through every last mile. Now I bet you’re remembering that warning I gave you at the top. Cheer up, I’m almost finished.
4. The lack of familiarity with electric vehicles in the general public tends to devalue them in comparison to their true value.
This is perhaps the most significant obstacle electric vehicles face, and the reason why I go to such lengths to describe their hidden benefits. A large part of what makes something valuable is how popular it is, and it can’t become more popular without becoming more familiar. It’s a chicken-or-egg problem. It is obviously true that a person can’t really know whether they prefer something without experiencing it…they don’t “like” it without at least knowing what it is. (Well on Facebook they might…)
What this boils down to is that even if you think you prefer gas cars over electric vehicles, you won’t ever really know until you at least know something about the latter. And if you’ve come this far on the post, you might as well go test drive one, too, because reading about it…well it makes you book smart and everyone knows it’s better to be street smart than book smart.
Even smartasses know that.
So since we already have a reasonably level-playing field at our disposal — albeit one with several generous assumptions on behalf of combustion cars — let’s now look at the optimizations a person could make if they wanted to try an electric vehicle. What they’d do if they really wanted to save every last penny they could, and not compromise their ability to drive wherever they pleased.
But before we do that, let’s take a break for some pictures…
With global EV sales having surpassed three million units as of November 2017, it’s very clear that people are already making them work — even if you just scrolled right past all the visual proof.
Here’s a few more things to help encourage you to try an electric car sometime soon. Hell, call Elon and tell him I said to give you a good deal on a Model 3. He probably won’t, and I don’t actually have his number, but it wouldn’t hurt to try.
And that’s the point, really. You really just have to try.
- The Federal tax incentive. Although I’ve left it out of the final calculations above, the Federal tax incentive of $7500 for the purchase of a new electric vehicle cannot reasonably be disregarded in a purchasing decision for a new car. With the full rebate still available from all manufacturers, applying it to the figures above would put the Nissan Leaf and the base Tesla Model 3 handily at the top of the pack, and the Chevy Bolt in a virtual tie with the Civic. The full rebate is available for six months past the end of the quarter during which a manufacturer passes 200,000 unit sales. After that, half the rebate ($3750 for electrics, $2000 for the Fusion Energi as shown in the table below) is available for the next six months, and finally one-quarter of the rebate ($1875/$1000) is available for six months, after which the rebate expires completely. One point of note, however — the vast majority of manufacturers haven’t even begun to significantly ramp up sales, so provided the incentive has not been cancelled, you will likely be able to get the full value depending on the car you want, as long as you have sufficient tax liability and decide to buy in the next two years or so.
2. Savings on solar. The possible long term savings from the installation of solar panels cannot be underestimated when used in conjunction with electric vehicles. There is a Federal investment tax credit on solar — as shown in Figure 18 below — which makes installing solar panels on your home an even more compelling investment for the next few years.
The cost of home solar has been dropping so quickly that now it is usually possible to get a low interest loan covering the installed costs (which vary by region but are roughly $2/watt installed after the ITC above.) What’s more, the cost savings reflected on your electrical bill is frequently enough to offset a typical solar loan payment — so it becomes a matter of whether your house is in a good location for panels, what kind of condition your roof is in, and whether you’re willing to invest the time to find out more about it. Tying this back to electric vehicle use — the greater the electrical energy requirements of your home, the more you will save over time by installing a photovoltaic system appropriate to those needs, because your per unit electricity prices will effectively be lower year round for the life of the system (typically 20+ years.) While system performance varies depending on where you live, a conservative estimate of $0.10/kWh per unit electricity over the life of your system would save you about $6,240 over 15 years driving a Leaf, $5,860 if you were driving a Bolt, $5375 if you were driving a RWD Model 3, and about $5,120 if you were driving an AWD Model 3.
Even in the solar-poor region of Rochester, N.Y. that amount of savings could buy you a 2.5 kW system capable of generating about 3,000 kWh/year. Such an upfront investment would pay for almost all the fuel you would need for driving the car!
To restate this in simpler terms, buying a 2.5 kW PV system for your home would generate over 90% of the energy necessary to meet the requirements of your electric car over its entire lifetime. That basically means that you can essentially “prepay” for your fuel by installing a solar system. If you offered Exxon $5000 for thirteen and a half years’ worth of fuel they’d call a paddy wagon and send you off to a room with no windows and a nurse to pass medication through a little slot in the door. You’d get free food but you wouldn’t be able to tell what it was and you’d have to eat it with plastic silverware.
3. It really depends a lot on what state you live in. No, I don’t mean a state of denial. I mean your actual state — and what your electricity versus gas prices there are. You might find the below table interesting. It highlights the states in which you’ll find an even greater advantage by driving an electric car: places where electricity doesn’t cost as much as the starting estimate above ($0.1301/kWh), and places where gas costs more than starting estimate above ($2.62/gallon.)
What all of this means is that if you choose to install solar panels and can find an electric car which suits your driving habits, you will have the least expensive car to operate that you can possibly have.
In fact, I’m so confident about that assertion, I defy anyone reading this to invent a reasonably rational situation in which it isn’t the case — and I’ll even go so far as to help you figure out how you’d go about the business of trying to invent one.
- You would have to begin your comparison against a combustion car with an absurdly high strict combustion fuel economy — like 50 mpg. Without doing that, you’d be forced to make even wilder assumptions than the ones I’ve listed below this line. In other words, you’d need a very high performing hybrid or a strict combustion car at the very top of its class. You’d basically be driving a Prius or…well, a Prius. (Okay, to be fair you could probably get away with a 4-cylinder, six speed manual 2018 Chevy Cruze diesel with an EPA estimated 52 mpg on the highway — but you’d have to stay away from city driving where its economy plummets to just 30 mpg — bringing it down to 37 mpg overall. Staying away from city driving as a general rule doesn’t seem ‘reasonably rational’ so let’s drop the Cruze lest we really be here all day.) So it’s a Prius or another hybrid of comparable fuel economy which doesn’t have nearly the reliability of a Prius, and hence reintroduces greater repair costs…okay, let’s base it off the Prius. The Prius has a startlingly low 15-year maintenance cost of about $9000 (at least if you don’t consider replacing its drive battery in years 11–15), so it checks that box too.
- You would have to live in an area where your solar insolation is very low, or your overall electricity consumption aside from your car’s requirements is very low, or both. Setting aside Alaska (where just 0.23% of Americans [1 in 435] live and where average solar insolation, at about 2.7 kWh/m2/day in Anchorage is less than 60% of the U.S. average) you would have to be a person who doesn’t use very much electricity. With the national average being about 10,700 kWh/year (low in Hawaii at roughly 60% of this figure) — you’d have to use about half as much as a typical American, and you’d have to assume a maximum of about 50% waking hour use (during which solar panels without a residential storage battery would be most effective.) This despite that most typically, more electricity is used during the day than at night time.
So to this point, you’d have to be a Prius driver with very low home energy requirements. Let’s now take a look at what your driving habits would need to be…
- You would need to assume the vehicle travels on average more than at least a minimum of 200 miles a day.
This would almost definitely make you either an UBER driver (<300,000, or <0.13% of Americans) or a “stretch commuter.” The U.S. Department of Transportation defines the latter as people who are willing to drive more than 50 miles one way to work. True, there are about 3.3 million people (1.48% of Americans) who drive this far, but only 13% of those people (433,000–0.188%) drive 100 miles or more one way. Because it’s a nonlinear drop off rate, we know that far less than 1.9 million (0.844%) people drive more than 75 miles one way to work every day, which you’d almost certainly have to do if you were going to average a total of 200 miles or more in total. So you’d wind up being a Prius driver with low home energy requirements who lived a long way from work. You’d also most likely be working in construction or manufacturing (or perhaps hold a managerial job), have a marginally lower chance of making your living driving for UBER, and slightly more than 5 out of 6 of you would be male.
No offense, but what sort of construction worker arrives on the job site driving a Prius (here comes the hate mail again…) and what sort of managerial job would pay so comparatively little that you’d be using fuel economy as the primary metric by which to choose your vehicle? [Author’s note: let me know if you caught that pun.] I mean all of these things are certainly possible…
So to this point, you’d probably be a stretch commuter driving a Prius over 75 miles one way to work who also has very low home energy requirements and who also drives another 92.8 miles per day on average for non-work related tasks — at least provided you work five days a week. That last requirement snuck in there because it’s necessary to keep the average at 200 miles or better. If you didn’t have that requirement too, you’d have to assume that you drove more than 75 miles one way to work daily, dropping you into a smaller group than just 0.974% of Americans.
Now I know what some of you are thinking.
You’re thinking, “But…but…well, but what about the standard deviation? I mean, let’s suppose I usually drive under a hundred miles a day but sometimes I drive, say, 500 miles. That blows your theory right out of the water, now doesn’t it?”
Sigh. You caught me.
I tried to slip one past you. The old “road trip” principle. The ace-in-the-hole. The Trump card. One of us must have missed a detail somewhere, I guess. Like I never would have considered the possibility of a road trip — who does those anymore? There’s nothing to see in this country anymore and Scott Pruitt is starting to consider making damn sure of that by cutting a few of these down:
Apparently they make excellent picnic tables and firewood — and as an added bonus, there could be a bit of coal & oil underneath them. We might as well party like it’s 1999, right?
Okay, but I wish you’d at least make up your mind. I mean, are we going to respect history,
…or aren’t we?
In fact, if you were paying close attention, I brought my own ace-in-the-hole. Those with sharp eyes can see it on the back of my car in some of the above photos. [Hint: it’s green and it’s a club. Teddy would have been proud, I think.]
My ace? It’s the fact that if you’re driving more than 500 miles, but you only need to do it every once in a while, and you’re trying to do it as fast as you possibly can, you’d bloody well either rent a gas-powered car on those occasions or you’d fly.
But…but…but…<searches pockets for more cards> but I want maximal convenience!
Still can’t make up your mind, eh? I thought we were talking about less expensive, now it’s only convenience that matters?
No, I want both! Dammit I want both, I’m an AMERICAN! I deserve both at the same time! With a cherry on top and chocolate sprinkles and we can have unicorns and rainbows in everyone’s back yard! It will be great. Believe me, the best! Our generation will be HUGE!
Alright, if you want less expensive AND more convenient and you rarely need to drive very far, you can probably get away with buying a reliable used car and keeping it garaged most of the time. The better the fuel economy the more often you’ll be able to drive it for the same money (um…doh) and that will probably mean the least expensive car you can find on the used market.
I bet you’ll never guess what sort of car that would be.
But supposing you don’t live near an airport, or you’re afraid of flying…
…maybe you like the smell of gas fumes, your best friend works at the gas station, or you’re considering suicide and the garage seems like a compelling place to do it…
Okay then, by all means. Buy another gas car. Really. It won’t be obsolete for the next ten years or so, anyway. The Koch brothers will both be dead by then, but you know their snot-nosed children will fight just as hard for your continuing ability to keep right on burning fossil fuel — even though it’s pretty clear you no longer need to, and financially unwise as well.
But…what about the battery? Battery degradation! That’s it! I caught you!
Wow. Holy hell are you persistent! You caught me AGAIN!
Alright, let’s now make an adjustment for anticipated battery degradation — since it’s true that range can become a factor — particularly when it drops below about 100 miles per day. By state, driving distances look like this:
The easy way to look at that chart is to combine each pair of blue and green bars and consider them a fraction of the whole bar per state — that’s the fraction of days in which users drive 100 miles or less.
The warranties on electric vehicle batteries range significantly, with a referenced GM warning here  that users could see a capacity loss of up to 40% in the warranty period of 8 years, 100,000 miles. Tesla’s battery warranties diverge slightly depending on the options — the company guarantees 70% capacity retained by the short range version of its Model 3 in the 100,000 mile warranty period, and a similar 70% capacity retention in 120,000 miles for the longer range version. The Leaf is a bit trickier, as its battery warranty covers degradation below “nine bars” of capacity — which creates confusion with owners, many of whom likely don’t know that the “capacity bars” it uses are nonlinear: specifically that the first bar is worth almost two and a half times as much (at 15%) as the remaining bars (at 6.25% each.) This yields a curious warranty to 66.25% of capacity as shown here [26.]
Assuming compounded linear degradation of each of these batteries (that is, consistent degradation year by year, by percentage), and using the above guarantees and the conservative EPA range estimates posted for the vehicles yields the following (assuming the worst performance possible without activating the guarantee):
[Note: the battery warranties for these vehicles, at this usage rate, only last until about 7.5 years have passed. However, since the premature degradation of the battery would activate the warranty, these figures should in some ways be looked at as the most pessimistic projection — because if you wind up having your battery replaced, you’ll effectively return to full capacity, lasting you another 7.5 years. At least provided the replacement comes with the same guarantee.
As the above table shows, it’s important to look at your real world use in conjunction with the economies of the vehicles; if you regularly drive more than 100 miles per day and you’re looking at the 2018 Leaf, you should compare its six-year ownership cost to the others, not its eleven or fifteen year cost (unless you have a second car, that is.) The converse is also true, however. Even more than with typical consumer products, you usually do not want to buy appreciably more range than you need. At least not with all else remaining equal. Most electric vehicle drivers find themselves using the car more often than they expected rather than less often — though if you don’t believe me there are enough of us around now that it shouldn’t be hard to find someone other than me to ask. Not only do people typically drive them more, but as my travels across the U.S. in an EPA-rated 75 mile 2013 Nissan Leaf indicate, you can find charging stations just about everywhere once you start looking.
So now this really is the end.
Are you willing to believe the truth, or are you ready to face the consequences?
I MEAN IT NOW! THIS IS THE END!
Seriously? You’re STILL here? You’re starting to annoy me a little. And if you think I lose my patience easily, then say hello to my little friend…
When THIS guy gets pissed off, stuff gets broken seriously.
Thanks to…in no particular order…
All the people who signed my car.
All my Indiegogo supporters.
All the people who offered a place to sleep, something to eat, good conversation, a fist bump, a thumbs up in traffic, or even a laugh. Even when that laugh was directed AT me rather than with me. Sometimes you gotta laugh at yourself, and it helped.
To my primary sponsor, KOA — without whose support I would not have been able to afford the trip, and especially to Peggy who pulled me through more than a couple sticky spots.
To Michelin, who sent me tires that lasted twice as long as the OEM radials. Tires which made it across the Rocky Mountains twice, Glorieta Pass in New Mexico, and onto Los Alamos, New Mexico without me even realizing they had two nails in one of them.
To the excellent people at Plug In America, the countless EV club members I visited, to the people who lavished unspeakably nice hotel rooms on me (you know who you are) and to the excellent folks at HelpYourNeighbor.com and ACE Consulting.
Finally, it would be ridiculous not to mention the much needed respite of the granny flat most generously provided by the Bollars, the gentleman who offered me my first test drive in a Model S
(so exhilarating it still shivers the spine typing that more than 28 months later, and made frightening by the pack of deer that crossed my path while I was behind the wheel), UBuyGas funnyman Sal Cameli of Cameli IT Services, and to the person who showed me this:
…not because it’s too fabulous a teapot, but because I can’t for the life of me figure out either his name or his wife’s name, and even though I know exactly where the cottage they offered me for the night was, I can’t pinpoint their address on Google Maps to send a more formal thanks.
If you enjoyed this content, could you please do at least one of the following:
- post it on your social media feed somewhere
- like it somewhere
- talk to your friends about it
- find me on www.Patreon.com as Brian Kent or search for Negative Carbon Roadtrip.
- Tweet out a hello to Brian Kent
- email your comments directly: email@example.com
- Find me on Facebook at https://www.facebook.com/TheNegativeCarbonUSRoadTrip/ and like or share that page.
Even just plant a tree somewhere, or smile like you know something no one else does. Because now you sorta do, eh?