On behalf of the german Green parliamentary group, a paper entitled „Comparing the lifetime green house gas emissions of electric cars with the emissions of cars using gasoline or diesel“ was published at the end of August 2020, which was extremely well received by experts in this field.:
- Prof. Claudia Kemfert, member of the German Advisory Council on the Environment, wrote on Twitter: „And another study that clearly shows electric cars cause significantly less CO₂ than previously assumed.“
- Volker Quaschning, Professor for Regenerative Energy Systems at the HTW Berlin, said (also on Twitter): „Once again a study proves that the climate balance of the E-Auto is much better than that of the petrol and diesel. After 30000 km the additional emissions of the battery set are balanced. When will the excuses finally stop and when will we start the turnaround„
- Christian Bauer, environmental systems scientist at the Paul Scherrer Institute, one of the largest technical research institutes in Switzerland, said in SPIEGEL: „The study is cleanly conducted and explains in a catchy way how great the advantages of an electric car already are today.”
The conclusion of the authors Hoekstra and Steinbuch on electric cars is (quote):
„The greenhouse gas savings for cars sold in 2020 range from 54% to 82%.“
We want to know if this is true.
On page 5 they present a table with the results. Three conventional cars are compared with an electric car of the same class:
|Toyota Prius 1.8l 2020||Volkswagen eGolf|
|Production without battery||28||24|
|Production of the battery||–||11 (36 kWh battery)|
|Total g CO2eq per km||168||78 (54% less)|
|Number of km the electric vehicle needs to „pay back“ the battery||28 000 km|
|Mercedes C 220d||Tesla Model 3|
|Production without battery||32||28|
|Production of the battery||–||23 (75 kWh battery)|
|Total g CO2eq per km||260||91 (65% less)|
|Number of km the electric vehicle needs to „pay back“ the battery||30 000 km|
|Bugatti Veyron||Porsche Taycan S|
|Production without battery||40||36|
|Production of the battery||–||28 (93 kWh battery)|
|Total g CO2eq per km||778||140 (82% less)|
|Number of km the electric vehicle needs to „pay back“ the battery||11 000 km|
Initially we will focus on driving emissions.
(„Hoekstra and Steinbuch“ will be abbreviated as H&S in the following):
|Model||Driving emissions according to H&S [g CO2eq/km]|
|Toyota Prius 1.8l 2020||140|
|Mercedes C 220d||228|
|Tesla Model 3||40|
|Luxury class car|
|Porsche Taycan S||76|
The relative improvements from electric cars are enormous:
|Vehicle class||Emission saving of the electric car|
However, such high values should make expert readers suspicious. For however efficient the electric drive may be, as long as half of the electricity is produced by fossil-fuelled power stations, a 90 % reduction in greenhouse gas emissions is not very plausible. Something may be wrong with these calculations.
An essential basis for determining emissions is the energy consumption of vehicles. H&S have given comprehensible reasons why they want to use the information from the EPA (the American Environment Protection Agency). But then a curiosity follows: if no value is available for a certain vehicle there, they use another source for only this one.
This step must be viewed critically. It would have been scientifically correct to use the same data source for every vehicle comparison. This would have meant selecting only vehicles for which fuel consumption values are stored there. This brings us to
Fault no. 1: Comparisons with consumption values from different sources with different methods of determination are meaningless
On page 5 it says: „The rest of this document explains the calculations in detail and provides sources.“
This promise is not kept:
- The assumed consumptions are not mentioned
- Spritmonitor.de and the EPA are mentioned as sources – but which consumption was taken from which source is not indicated (it can only be deduced indirectly from the text)
- There is not even a record of what fuel the cars run on
The critical reader is expected to look elsewhere for the emission factors used by H&S and thus to draw conclusions from the absolute emissions to the consumption per 100 km. Such carelessness makes it difficult to check the plausibility of the data:
Fault no. 2: H&S try to disguise the origin of their values
So let’s make up for what the authors have missed. Regarding the assumed emission factors, the text reads:
- Page 21: „Emissions per litre are thus 3310 g for diesel and 3140 g for gasoline.„
- Page 14: „All in all electric vehicles sold in Europe in 2020 should count on 250 g CO2eq/kWh electricity over their lifetime.“
This value takes into account 31% of losses on the way of electricity from generation to the vehicle (see page 16).
The fuel and electricity consumption reconstructed by backward calculation from the reported emissions are given in column 3 of the following table. In addition, the fuel consumption of these vehicles according to Spritmonitor.de and EPA is also shown:
|Model||H&S-Driving emissions [g CO2eq/km]||(Calculated) H&S consumption [l/100 km or kWh/100 km]||Spritmonitor-consumption||EPA-consumption|
|Toyota Prius 1.8l 2020||140||4,46 B||4,44|
(Year of manufacture: 2018 to 2020)
|4,52 (= 52 mpg)|
|Volkswagen eGolf||43||17,2 E||15,0||18,6 (=30 kWh/100 ml)|
(Year of manufacture:
|40||16 E||18,3||16,8 (=27 mpg for the Midrange-Modell)|
|Bugatti Veyron||738||23,5 B||23,5 B||23,5 B|
|Porsche Taycan S||76||30,4 E||Not available||30,5 (49 mpg)|
(Source of EPA consumption: https://www.fueleconomy.gov/feg/bymodel/bymakemodelNF.shtml)
The evaluation of this table reveals further inaccuracies and inconsistencies: H&S claim to use EPA values where available. But this seems to be the case only for the two luxury cars. In all other cases there are discrepancies (which are not commented by the authors).
The value for the diesel car is said to have been taken from spritmonitor.de – but cannot be verified either: Mercedes C220d from model year 2019 on consume on average 0.26 litres less than the H&S emissions. This is the
Fault no. 3: The input data are unreliable. Four of the six consumptions cannot be proven with the sources given.
Another strange thing is noticeable: How does the Bugatti Veyron get into the list?
The goal of this comparison is stated as being to determine „the GHG savings for calls sold in 2020“ (accurate quote). However, production of the Bugatti ended as early as 2015; the authors are thus ignoring the technical progress of the luxury class over the last ten years or so. In addition, the total number of units sold worldwide was only 450, and hardly any German is likely to encounter this vehicle.
Wouldn’t it make more sense to compare the electrified Porsche Taycan S with another model that was produced in larger quantities? The Panamera was available in August 2020 with engine outputs ranging from 330 to 680 hp. Spritmonitor.de lists 12.3 litres as the average consumption over all years of manufacture. The EPA even states that the model with petrol engine consumes only 10.7 litres. Could this be the reason why the authors had decided on the Bugatti Veyron – the extremely high consumption of 23.5 litres/100 km? After all, this was the only way they were able to announce an alleged lead of the upper class electric car of a spectacular 82 percent. This manipulation is
Fault no. 4: The selection of the luxury car is not based on the vehicle market, but on the surprise effect on the reader
Going further with emissions.
Emission factors per litre of fuel are claimed to be „3310 g for diesel and 3140 g for gasoline.“ These are unusually high values for which no citable source is given. Instead, a new surcharge on combustion emissions is defined ad hoc. Without mentioning the source, the authors cite sections from a previous article by Auke Hoekstra. Especially this section (not quoted in the current paper) is quite informative (translated from German into English):
„I began to search dozens of scientific sources. But man, it’s hard. It cost me my whole weekend! The oil business is a world of its own, with different priorities, sources and even different units. To make matters worse, there are big differences between oil sources and refineries. And all this was very hard to read for a non-expert like me.“
There is no known publication of this text in a scientific context.
Furthermore, there are reasons to believe that Auka Hoekstra did not work too carefully during this weekend. Even the initial values for emissions without provisioning costs are wrong. Quote: „So this is in addition to the 2420 gr/l for pure gasoline and 2670 gr/l for pure diesel.“
These figures cannot be found in the source cited. Only „emission factors“ in kg/GJ and „caloric values“ in MJ/kg are given there. The deviations could be due to errors in conversion to another unit of measure (litres instead of kilograms).
More reliable values could have been adopted by the authors, for example from the Bavarian State Office for the Environment, Wikipedia or the Helmholtz Institute. For gasoline 2,330 to 2,370, for diesel fuel 2,630 to 2,650 grams/l are given there.
It therefore seems advisable to use reliable sources with generally accepted values for the emissions including upstream chains:
- In October 2019 the Austrian Federal Environment Agency reported 2,723 g/l for gasoline and 3,098 g/l for diesel fuel.
- Those who are interested in higher values can refer to a table of the Bavarian State Office for the Environment, for example. There, 2,877 grams are given for gasoline and 3,156 grams for diesel.
Fault no. 5: The authors use unusually high fuel emission factors of dubious origin
In contrast, emissions from electricity production are set at an exceptionally low level of 250 g CO2eq/kWh. The reason for this is a cardinal error of this paper – the application of the average electricity mix.
„We could look at the so-called merit order and see which is the energy source with the highest marginal cost (often coal) that would be switched off if electric vehicles were switched off. But why should we do this for electric vehicles and not for the heat pump or the TV or the air conditioning or the new factory?“
This paragraph is more than remarkable. For by making the factually correct statement that the energy source with the highest marginal costs (often coal) would be switched off if electric vehicles were switched off, Hoekstra & Steinbuch explicitly acknowledge that the additional electricity for electric cars is largely produced by coal-fired power plants.
The question why this logic should not also be applied to climate balances of other electricity-consuming products is easy to answer: the reason is the purpose of this life cycle analysis.
In the case of televisions (or, for example, refrigerators), the focus is on established products that have been in use for some time and will continue to be so in the foreseeable future. Since no decision has to be made on whether to buy or get rid of them, such product segments are not the focus of a discussion on measures to combat climate change.
The situation is different for new and additional products – especially those that are to be introduced solely and exclusively for climate protection reasons and whose suitability for the purpose has yet to be verified, such as the electric car or the heat pump.
Political decision-makers basically have the option of promoting new electricity consumers on the basis of environmental impact or to impose negative sanctions. In order to be able to make a meaningful decision, they need to know what impact the introduction of only this new product will have on greenhouse gas emissions. This cannot be determined on the basis of the average electricity mix, as this also includes green electricity. Eco-electricity is always produced and consumed – even if no new electricity consumers are added. Of course, the only relevant factor for the climate balance is which power plants supply the additional energy that is only needed for the new product. These are exclusively controllable, i.e. fossil-fuelled power plants.
The quotation at the beginning of this section shows that Hoekstra & Steinbuch are well aware of these relationships.
Fault no. 6: Hoekstra and Steinbuch try to conceal with rhetorical questions that electric cars are additional electricity consumers, forcing the production of fossil electricity to be increased by the charging current
H & S increase this error far beyond what is known from other studies by using an implausibly clean power mix.
The German fossil electricity mix currently has emissions of 860 grams when averaged according to the share of generation.[i] A scientifically correct study would have to set at least this value for the generation of charging current.
Most greenwashing studies apply the German electricity mix, which in 2019 had emissions of 410 grams/kWh. This means that they can understate the emissions of electric cars by a factor of 2.1.
Although the study was commissioned by a parliamentary group in the German Bundestag, H&S does not use the German, but a European electricity mix (self-defined including future improvements) with only 250 grams/kWh.
Fault no. 7: To determine the greenhouse gas emissions caused by electric cars, H&S uses a charge current mix that is too clean by a factor of 3.4
H&S also write: „The assumption that electric vehicles run on coal therefore only works if you disregard the development of the mix over the lifetime of the vehicle.“
This ties in with the widespread notion that charging electricity will become cleaner over time thanks to the increasing share of green electricity. But this is not true.
The emissions of additional consumers (such as electric cars) do not depend on the average electricity mix, but on the additional electricity mix. The following graph illustrates this relationship:
Principle of the merit order; source: Wikimedia Commons
(Kernenergie: nuclear power; Braunkohle: lignite; Steinkohle: hard coal; GuD: gas and steam combined cycle power plants; Gasturbinen: gas turbines; Heizöl: fuel oil; Grenzkosten: marginal costs; Ladestrom: charging current; Residuallast: residual load)
The „residual load“ is the controllable output of the power plants with non-renewable energy sources. Only these can adapt their production to the electricity demand. Green power plants always feed as much electricity as they can; their output fluctuates greatly, and they are basically not controllable.[ii] Electric cars are new, additional electricity consumers. Their carbon footprint therefore depends solely on fossil electricity, which would not have to be generated if the electric cars were not used. Thus, for forecasts of future emissions, only the change in the composition of the fossil power mix over time is relevant.
Fault no. 8: H&S try to hide the fact that the increase of the green electricity quota cannot improve the climate balance of the electric car as long as the marginal electricity is generated with fossil energies
For a revision of the H&S values, the power losses depending on the data source still have to be determined. (By the way, this is the only point in the H&S paper that deserves praise. For the losses from electricity generation to the socket at the consumer’s end, a quite realistic 31 percent is given, citing a European study).
The total losses are made up of three components:
- The power plants‘ own demand for electricity
Gross electricity generation in 2019 was approximately 607 billion kWh, while net electricity generation was only 518 billion kWh; this represents a 17 percent increase on net electricity.
- Transmission and distribution losses
Westnetz, the largest German distribution system operator, reported losses of 7.17% for 2019. However, this is in addition to the losses in the upstream transmission network. Ten percent as total losses is therefore a conservative assumption.
- Charging losses
In 2020 the ADAC determined values of 10 to 20 percent for most electric cars. For the European bestseller Renault ZOE it was 19 percent, for the Tesla 3 even 24.9 percent. We choose a cautious 12.7 percent.
The individual losses must be multiplied by each other: 1,1 x 1,17 x 1,127 = 1,45
The total losses thus amount to approximately 45 percent.
Now all necessary data are available to replace the results of H&S by a realistic evaluation. In contrast to H&S, we use the same source for consumption in every vehicle class throughout.
Power consumption of spritmonitor.de does not include losses due to rectification and charging management, so it has to be increased by 45 percent. EPA values include charging losses, here the surcharge is 29 percent according to the above calculation.
As fuel emissions we use the relatively high values of the Bavarian State Office for Environmental Protection: Gasoline (incl. upstream): 2.877 g/l; Diesel (incl. upstream): 3.156 g/l.
We estimate the fossil current at 860 grams/kWh (see end note i).
We replace the Bugatti, which is senseless for this comparison, with the Porsche Panamera with gasoline engine. Here are the results:
|Model||(Calculated) H&S consumption [l/100 km or kWh/100 km]||Spritmonitor-consumption [l/100 km bzw.|
|H&S-Driving emissions [g CO2eq/km]||Realistic Driving emissions: [g CO2eq/km]|
|Toyota Prius 1.8l 2020||4,46 B||4,44||140||128|
|Volkswagen eGolf||17,2 E||15,0||43||187|
|Mercedes C 220d||6,88 D||6,62||228||209|
|Tesla Model 3||16 E||18,3||40||228|
|Bugatti Veyron||23,5 B||Not available||738||676|
|Porsche Panamera||–||10,7 B (EPA)[iii]||–||308|
|Porsche Taycan S||30,4 E (EPA)||Not available||76||336|
The following are the differences in driving emissions (electric car minus combustion engine) in relation to combustion engine emissions in percent:
|Vehicle category||H&S claim that the electric car has lower emissions by||After correction of the errors it has increased emissions by|
|Compact||-69 %||+46 %|
|Median||-82 %||+9 %|
|Luxury||-90 %||+9 %|
A bar chart visualises the greenwashing effect of the assumptions made by H&S:
Electric cars perform worse in all three vehicle classes – sometimes more, sometimes less; they will never „pay back“.
The use of higher emission factors (e.g. based on new research results) would at best lead to approximately the same emission level depending on the vehicle model. Highly efficient combustion engine cars such as the Toyota Prius (as a non-plug hybrid) would continue to have a climate benefit.
The corrected values also relativise the importance of the battery pack for the climate balance. For if the electric car performs worse than the combustion engine simply because of its driving emissions, it is irrelevant whether an attempt was made to gloss over the emissions from battery production with average current.
Outlook for the future
The current disastrous emissions from electric cars due to the fossil energy mix will improve significantly when coal-fired power plants are shut down in 2038. These will be replaced by gas-fired power plants. However, at around 490 grams/kWh, emissions will still be about twice as high as Hoekstra and Steinbuch claim for 2020.
Even then, the electric car will not have a climate advantage, at least not over natural gas cars. Because if both types of drive use the same primary energy, the basic advantage of the electric car, i.e. lower energy consumption, will be offset by the high electricity losses between production and consumption (about 45 percent in total).
On the last pages H&S allow themselves another serious mistake. Quote: „Carbon dioxide emissions are a fundamental characteristic of combustion engines that will never disappear.“
This sentence is likely to mislead the reader. It gives the impression that, in contrast to other vehicles, only electric cars could drive emission-free in the future.
Indeed, for a long time to come, e-cars will continue to cause emissions during battery manufacture and electricity generation.
Internal combustion engines, in turn, can run on synthetic fuels produced from renewable sources. They then achieve emission levels roughly equal to those of e-cars (see this plausible estimate).
Such fuels will be an indispensable element of future electricity-based energy supply. In its report „Greenhouse gas-neutral Germany in 2050“ , the UBA estimates net electricity generation for 2050 at 3,000 TWh, about five times as much as before. This will require large quantities of chemically bound energy to be imported – e.g. as synthetic fuel. The necessary infrastructure will have to be created in any case. It would therefore be consistent and far-sighted to take this into account today when planning future motorised private transport.
Fault no. 9: H&S try to prevent an open-technology implementation of the energy turnaround with demonstrably false assertions about combustion engines
This paper is an illustrative example of how the application of the average electricity mix, although not objectively justifiable, offers the simplest, most effective and therefore most widespread way to produce greenwashing studies.
In an earlier article (quoted above) Auke Hoekstra said:
„I have been writing for about fifteen years about the inevitability of electric vehicles. Sometimes people call me an evangelist…“
Auke Hoekstra is unfortunately not alone in this. It is precisely this mentality that has led to a flood of systematically incorrectly conducted studies, which are now intended to legitimise state coercive measures to enforce electric mobility. The combustion engine industry is threatened with extinction, even though electric cars are demonstrably unable to reduce greenhouse gas emissions. Germany in particular will have to struggle with the consequences of this misguided approach in the coming years, as the following two news show:
News from China: „As in Europe, the combustion engine was an object of hatred in China. They have even closed down corresponding university institutes. Now they are supporting the development of new internal combustion engines that have a high compression ratio and already have efficiency levels of 50 percent, and they are using synthetic fuels. This makes the combustion engine absolutely clean. Fortunately, the Austrian company AVL is involved.“
What will remain of the former know-how of European engine developers when the political elite finally realises that electric cars cannot replace the internal combustion engine?
Kai Ruhsert, October 9, 2020
[i] If there is no possibility of a mathematical simulation with a resolution in minute intervals, a reasonable estimate must be attempted, e.g. like this: The emissions of the primary energy sources including upstream emissions (always in the order lignite, hard coal, natural gas) amount to 410, 390, 230 g CO2/kWh (source; they refer to IFEU and UBA)
Gross power plant efficiencies in 2019 were 0.394, 0.443 and 0.469 (source: UBA). That results in: 1041, 880 und 490 g CO2/kWh.
According to energy-charts.de, lignite, hard coal and gas contributed 102, 50 and 53 TWh to electricity production in 2019. This results in average fossil emissions of 860 grams. (As coal moves to the right in the merit order due to rising CO2 certificate prices, this value is probably still far too low.)
[ii] The fact that a lot of wind and sun can lead to negative electricity prices on the electricity exchange is an economic, but not an ecological problem. Green electricity always pushes fossil electricity out of the grid, even if it is delivered abroad at very low prices.
[iii] Neither for the Bugatti Veyron nor for the Porsche Taycan fuel consumption values from Spritmonitor are available. As the emissions of the Taycan can only be calculated with the EPA consumption, the EPA value was also used for the Porsche Panamera. This makes it possible to take the data for the two luxury class vehicles from the same source.
The EPA consumption takes charging losses into account. The Taycan’s electricity consumption was therefore only increased by a factor of 1.17 (internal consumption of the power plants) times 1.1 (transmission and distribution losses).
Spritmonitor-power consumption figures do not include charging losses and have therefore been increased by 45 percent.