Review of „How clean are electric cars? T&E’ s analysis of electric car life cycle CO₂ emissions“ (April 2020)
(A german version is available here)
The European Federation for Transport and Environment, commonly referred to as Transport & Environment (T&E), is a European umbrella for some non-governmental organisations. In April of 2020 they published a new study (all bold representations were added later): 
To bring clarity and transparency to this debate, T&E has produced a comprehensive and forward-looking comparison of electric, diesel and petrol engines in different car sizes for 2020 and 2030. The online tool , published alongside this paper, is based on the latest evidence that shows that an average EU electric car is already close to three times better than an equivalent conventional car today. Crucially, electric cars will get considerably cleaner in the next few years as the EU economy decarbonises, with average EVs more than four times cleaner than conventional equivalents in 2030.
Three or four times better than conventional cars?! That would really be amazing. Most electric car-friendly studies confirm only a double-digit percentage reduction. We want to investigate how this result could have come about.
As is well known, the climate balance of the electric car depends crucially on the greenhouse gas emissions attributable to the generation of charging current. Which electricity mix did they use? The following quotations allow an answer (boldface was added later):
„T&E’s calculation of the carbon intensity of a country’s electricity grid is done on the basis of a bottom-up calculation based on realistic electricity generation mix evolution …„
„… the calculated EU27 average carbon intensity of the electricity grid is 319 gCO₂e/kWh in 2020 and 168 gCO₂e/kWh in 2030 and 84 gCO₂e/kWh in 2040.„
„To model the decarbonisation of the electricity over the lifetime of the EV and the future trends in EV’s impacts, electricity production mixes in EU countries are based on …„
„… electric cars – powered with the average electricity …„
The study is obviously based on the average current mix. This choice inevitably has a strong impact on the results:
Short term marginal mixes are generally higher than the average. 
How does T&E justify their decision?
First of all, it can be said that they demand compliance with scientific standards by writing this:
New LCAs should be carefully peer-reviewed and should be assessed with regard to the added value they provide, the amount of new primary research that is included and to what extent the data used is outdated with regards to technological progress and latest scientific evidence.
So let us check whether Transport & Environment itself followed the scientific rules by looking into a generally accepted manual:
2010 the Joint Research Centre of the European Commission issued a General guide for Life Cycle Assessment: 
Until today, no commonly accepted guidance exists that would complement the general framework provided by ISO 14040 and 14044:2006. The ILCD  has been developed to fill this gap as decision makers in government, public administration and business rely on consistent and quality-assured life cycle data and robust assessments in context of Sustainable Consumption and Production.
The manual explains which methodological faults should be avoided when choosing the electricity mix:
Failing to choose the proper LCI  modelling principle and associated approaches
The choice of modelling principle – attributional or consequential – decides whether the technologies to be covered by the collected unit process data in the inventory analysis should reflect the average technology for a given region and time-period or rather the marginal technology that is increased or decreased in use as consequence of the studied decision.
The decision-context of the goal of the LCA determines the appropriate LCI modelling principle and method approach to be applied. Considering other issues such as reproducibility and robustness the practical guidance of this guidance document was derived.
The difference between the marginal and the average can be large for some technologies. In the case of electricity generation, the marginal technology can be coal-fired power or wind-power, while the average technology will typically look very different.
For products or systems that use much electricity, the single choice of electricity technology (mix) will often be decisive for the overall results and the wrong choice of modelling principle will then give misleading results. The same issue applies to all kinds of processes and is hence one of the most outstanding methodological choices in LCA.
To enable the reader to identify the appropriate methodology, several case studies are explained in detail:
Situation A: „Micro-level decision support“
Situation B: „Meso/macro-level decision support“
Situation C: „Accounting“
It is precisely derived under which circumstances marginal processes as well as „short-term“ and „long-term“ marginal effects have to be examined in order to arrive at meaningful results. The term „marginal“ is is not mentioned less than 111 times in the manual.
Believe it or not, the T&E paper does not contain the term „marginal“ even once!
How to identify the appropriate electricity mix
Had the author looked into the handbook, he would have found decision guidance:
In the example of electricity procurement, a consequential modelling would require the use of the mix of marginal technologies to be used. If – as in the example of hydropower – the specific procured technology hydropower is not scalable in production (as e.g. in Germany), the consequential demand for electricity is not resulting in additional hydropower installed but this is only resulting in a virtual shifting of electrons from the electricity market mix to the specific supplier. Using hydropower data would substantially change the results, while not being justified in the decision-making context of Situations A and B.
Just like hydropower in this example, green power generation in most European countries cannot respond to higher current demand. Although renewable energy capacity is being expanded, this is not a response to the energy demand of electric cars. The expansion is already planned anyway and would take place even if no electric cars are sold at all. To prevent erroneous results, greenhouse gas emissions in all scenarios, whether for today or 2050, must be calculated using the short-term marginal mix.
Not only did they completely fail to justify their methodology. They made the wrong choice too.
In January 2018, the German Federation of Energy Consumers named the core of the problem: 
But why shouldn’t the German electricity mix be used as the basis for calculating emissions? The explanation is simple, but it cannot be refuted: Photovoltaic plants, wind turbines and nuclear power plants always produce when they can and are ready for operation, as their variable costs are very low or almost zero. Because of an additional electric vehicle they will not (can not) produce more electricity. So the additional electricity has to come from a power plant that is not yet fully utilized. It would also be conceivable that less electricity is exported – but then a fossil-fuelled power plant would be started up abroad, because the same business principle applies to electricity generation there. … This situation will still prevail even if in ten or twenty years‘ time more than 50 percent of electricity generation is based on renewable energy sources. So anyone who reckons with the low emission values of the power plant mix is lying through their teeth – some consciously, some out of ignorance.
Choosing the right electricity mix helps to prevent the calculated multiple use of electricity from renewable energies.
Anyone who – like T&E – calculates the climate balance of an electric car using the average electricity mix acts as if electricity from renewable energies can be used twice: Once by the existing electricity consumers and then again by the new electric cars. If you grant electric cars even one watt of green electricity, you have to extract this energy from other consumers.
This is the reason why the average electricity mix leads to completely nonsensical results.
However, if the aim is really to gloss over the carbon footprint of the electric car, then the average electricity mix is of great advantage because of the greenwashing effect, as the following diagram illustrates (including upstream emissions):
The fossil electricity mix in Germany has emissions of around 874 g CO2/kWh, while the average electricity mix in 2016 was only 527 grams according to the Federal Environment Agency. With this lower value, the electric car (Renault ZOE) can be presented in a much more favourable light.
The Umwelt- und Prognose-Institut e.V. confirmed this finding in 2019: :
However, when realistically considering the marginal costs of using coal-fired electricity, the CO2 emissions of electric vehicles are significantly higher than those of petrol and Diesel cars.
No wonder T&E added another common mistake:
T&E has calculated the carbon footprint of the battery production based on the
carbon intensity of the electricity used for manufacturing processes.
In reality, the electricity needed for battery production is also additional demand that requires additional fossil electricity. This results in much higher CO2 emissions than T&E claims.
The false assumptions lead to false conclusions:
Renewable electricity generation is expected to increase from 35% in 2019 to 43% in 2025, 55% in 2030 and 74% in 2040 (see Figure 1). This rapid uptake in renewables has an important consequence: an EV bought today will keep getting cleaner throughout its lifetime, which is usually not considered in LCAs.
There are good reasons why this is not considered in serious studies. Their authors know that as long as fossil power plants supply the marginal electricity only the improvement of the fossil electricity mix may be included in the climate balances.
No matter how weak the foundation of T&E’s argumentation may be, the claim is all the greater:
The potential of electric cars to mitigate CO₂ emissions is crystal clear: on average EVs are close to three times cleaner than diesel and petrol cars today. Discussing whether or not coal-fuelled electric cars are better or worse for the climate than conventional cars is no longer relevant (EVs are 30% cleaner even then). The urgency should be placed on accelerating the transition to electric mobility while at the same time decarbonising the electricity grid.
They stubbornly stick to the wrong power mix:
To give an example, the calculated EU27 average carbon intensity of the electricity grid is 319 gCO₂e/kWh in 2020 and 168 gCO₂e/kWh in 2030 and 84 gCO₂e/kWh in 2040. Most studies do not rely on such dynamic data and rather use static, constant values for the carbon intensity of the electricity grid.
With such major conceptual flaws some mistakes on a small scale are no longer surprising. For example, they write about the charging losses of electric cars:
On top of this, 10% efficiency losses were added: 5% from the charger equipment and 5% from the battery charging efficiency.
If we look at the sources given by T&E itself, it turns out that the AC/DC conversion losses (5 %) were „forgotten”. The quoted source in turn disregarded the discharge losses. Therefore these claims also are far from reality. New vehicles reach approx. 18 % under favourable conditions, but with older batteries and in warm or cold weather the losses increase (for the influence of the operating conditions see also: „Battery electric vehicles in practice“).
Against this background, the final statement is more declamatory than convincing:
Awaiting for the grid to decarbonize before shifting to zero emission mobility would increase further CO₂ emissions and would seriously compromise any chances of reaching the Paris Agreement.
The contrary is the case, as the independent Energy consultant Mario Sedlak noted: 
Nine wind power plants save 35,000 tons of carbon dioxide per year. 20,000 electric cars cancel out these savings because they consume the electricity that would otherwise have replaced the production of conventional power plants … Switching from cars to equivalent electric cars is only a climate protection measure when the last conventional power plants have been shut down.
According to their summary T&E wanted „to bring clarity and transparency“ to the debate.
This publication, however, is nothing more than scientifically worthless advertising for electric cars.
Kai Ruhsert, June 2020
 ILCD = International Reference Life Cycle Data System
 LCI = Life Cycle Inventory
 https://www.energieverbraucher.de/de/energiewende__1900/NewsDetail__17550/; translated into English