Friday, August 6, 2021



Contrary to recent claims from a group of our MPs, EV demands on our power system infrastructure do not lead to national bankruptcy.

The All Party Parliamentary Group opposing government policy on electric vehicles claim in a new report that the required investment in electricity generation will “bankrupt UK plc”. Unfortunately for this claim, it is based on some major errors of fact and understanding. These exaggerate investment costs by a factor of 25 or more, even on the basis of a rather pessimistic cost assumption that is the authors’ starting point.

Thirteen MPs and Lord Lilley have endorsed a new report[1] from the All Party Parliamentary Group[2] for Fair Fuel for UK motorists and UK hauliers. The press release describes it as ground-breaking. However the report consists largely of a stitching together of questionable “facts”, opinions and arguments from multiple sources and special interest groups, and contains major and consequential errors, of both fact and understanding, for electricity in particular.

Inevitably a central focus of the analysis, and for press headlines, is the additional investment cost required to service the additional electric vehicles implied by government targets. It is this estimate of required investment cost that has provoked the “bankruptcy” claim and headline, so it is worth examining its credibility. The  calculations are set out quite unambiguously in the report.

“FairFuel UK and the ABD have also analysed the economic consequences of dumping ICEs and concur with ... [the] prodigious cost conclusions shown here, to all our lives. The Government’s unilateral decision to commit UK to 75% reduction in carbon by 2030-2035 will increase our national debt by £2.17 trillion.[3]

2020 Consumption of petrol, diesel, bio diesel, bioethanol was 30 million metric tonnes per year. … A 75% reduction target by 2030 -2035 or whatever target date is chosen, means finding enough energy to replace 22.5 million tonnes of fuel per year.

The average energy density of the fuels: petrol, diesel, bio diesel, bioethanol is 12.5 kWh per kilo. That [equates to] 281.25 TWh pa.

Considering one nuclear power station generates about 4 TWh/year, [this] means 70 nuclear power stations are required to offset the 75% reduction in petrol and diesel.

A nuclear power station costs £22 billion, so 70 nuclear power stations =  £1.54 trillion.”

Order of Magnitude Error 1. The calculation ignores efficiency in use.

What matters to the consumer, and hence for aggregate energy consumption, is the efficiency with which different fuels can provide useful energy to deliver the service required.  Internal combustion engines (ICE) have very low efficiencies in the delivery of useful energy compared to electric vehicles (EVs), but the report assumes they are the same.

A typical claim made for EVs by charging networks is that EVs are 85-90% efficient while internal combustion engine cars 17-21%. If accurate, this implies we should divide the APPG estimate of 281.25 TWh by a factor of 4 or 5.

Calculating an equivalent electricity requirement is not just a matter of comparing technical parameters. The reality will depend on many factors including speed, driver behaviour and driving conditions, so there is no single universal answer. Hence it is worth comparing alternative sources and informed assessments. The following quote is from the Australian Energy Council[4]

“EVs convert over 77 per cent of the electrical energy from the grid to power at the wheels. Conventional gasoline vehicles only convert about 12 per cent – 30 per cent of the energy stored in gasoline to power at the wheels,” according to the US Department of Energy. …. Particularly in city driving, IC engines waste fuel while idling or operating at very low outputs compared to their design capacity, and engines at low output achieve very low efficiencies. … And, unlike EVs, most conventional vehicles do not recover the energy wasted to heat by braking for traffic lights.

The well-researched Committee on Climate Change (in its Sixth Carbon Budget[5]) uses an efficiency multiple of about 3, less than the above, but recognises that the nature of EV load also implies a less than proportionate requirement for additional capacity, a view shared by the National Grid.

Carbon Commentary also provide a first approximation estimate[6] of total requirements, for cars, based on an intuitively reasonable calculation from official transport statistics.

A 2017 electric car will typically get 4 miles from a kilowatt hour of energy.[NB 40 mpg for a typical petrol driven car would be equivalent to about one mile per kWh] The average car in the UK travels about 8,000 miles a year. That means that a typical electric car will use about 2,000 kWh a year.  In 2016 there were 36.7 million cars on the road in the UK. The total amount of energy required to power these cars if they were all electric would be about 75 TWh a year.  The total consumption of electricity in the UK last year was about 300 TWh. So if all car and taxi transport was by electric vehicles, the total amount of electricity needed would rise by approximately 20%.

This of course is an estimate that covers only cars and taxis, but assumes 100% rather than 75% replacement. The National Grid[7] likewise does not seem to be unduly concerned with the issue and suggests that “even if the impossible happened and we all switched to EVs overnight”, peak demand[8] (measured in GW not TWh) would only increase by around 10 per cent, or about 6 GW. One reason for this rather low number for peak capacity requirement is that vehicle charging load can be managed so that incremental TWh can potentially be met by much less incremental capacity than would be required for other consumer loads.

Order of Magnitude Error 2.

The report is also at odds with reality in its depiction of the scale of output expected from a 3.2 GW nuclear plant. It takes as its cost benchmark the reported numbers projected for the construction cost of EdF’s 3.2 GW Hinkly Point C nuclear plant (over £ 20 billion), but assumes an electrical output of 4 TWh for that plant. Hence it deduces the need for 70 Hinkly Point C equivalents, or 224 GW of additional capacity. However EdF have quoted an annual output (assuming 90% load factor) of 25 TWh, a scale confirmed by the government in its evidence to the Public Accounts Committee.[9] A more accurate statement of the expected output from this benchmark 3.2 GW plant therefore reduces the scale of investment by a factor of 6.

The error factors are multiplicative, and the result is that the implied 224 GW of capacity calculated in the report, for a 75% switch to EVs, exceeds the more realistic 6 GW suggested by National Grid, for a 100% changeover, by a factor of 37.

Is the Hinkly Point nuclear plant a reliable cost benchmark anyway?

There are several reasons to suppose it is not. The Public Accounts Committee was critical of government procurement performance, and clearly feels the cost of this contract was excessive. It is also generally assumed that subsequent nuclear power plant of similar design would be cheaper. Not least, not everyone will agree with the assumption that all-nuclear is the least cost route to expanding power generation, the implicit assumption that underpins the report’s cost estimates.

Dieter Helm[10] and others have also argued that about 50% of the estimated very high cost of Hinkly C is entirely attributable to the very high return to be earned by EdF over the life of the project. If this were indeed to be financed through addition to the national debt, one might surely expect to apply a much lower rate of return, closer to government borrowing rates. Helm argues for rates as low as 2 or 3%, halving the cost of a station such as Hinkly C. By implication Helm’s arguments alone would imply a further cost reduction by a factor of 2.


The misunderstandings in the report, at least on relative efficiency and likely cost, and an overall error factor of perhaps between 25 and 50, even on the basis of its own assumptions and methodology, are all the more surprising given that the House of Commons Library has just published (June 2021) an analysis, Electric Vehicles and Infrastructure[11], on the same subject. This is a brief but well-researched source of basic information on the subject under discussion. One might assume it was available to MPs.

In reality, if we do proceed successfully with low carbon generation, electric vehicles have a major positive role to play in helping to balance power systems associated with less flexible generation from nuclear or renewable plant. This makes them an economically net positive option in any low carbon future. But that is a bigger subject to which we can return, and which I have addressed elsewhere[12].


Updated on 3 September 2021 to include additional sources and references

[2] Readers should note that these informal groups of MPs do not have the official standing of parliamentary Select Committees.

[3] This number seems to be the national debt in April 2021, as referenced later. It is intended, one assumes, to be the £1.54 trillion calculated by the APPG researcher.

[8] In large power systems, it is important to distinguish between the additional energy requirement, kWh or TWh, and the amount of additional generation capacity required. The relationship between the two, for different types of consumption, is a crucial element of system economics.

[9] Hinkley Point C - Committee of Public Accounts - House of Commons ( The Department (BEIS) is recorded as stating that Hinkly Point C is expected to supply about 7% of UK requirements, ie about 20-25 TWh.

[10] Energy Policy: What happens next? - Dieter Helm

[11] CBP-7480.pdf ( Electric vehicles and infrastructure June 2021

 [12] Enabling Efficient Networks For Low Carbon Futures | The ETI


ADDENDUM. 19 October 2021

This APPG report has been modified since the above was posted, but the fundamental errors remain.

The relevant paragraphs have been revised to suggest that the output of a nuclear plant of the same scale as Hinkly C could be between 8 (not 4) and 25 TWh, and therefore that the EV requirement is for “between 12 and 35 new nuclear power stations” (not the more realistic 3 that use of more conventional arithmetic would imply).




Estimated energy requirement (c 4x factor error)

281 TWh

281 TWh

Output of “a Hinkly C” – a 3.2 GW station


8 -25 TWh

Number of such stations required



Cost of a station (presumably a 3.2 GW station)

£ 22 bn

£ 10-50 bn


So what has happened.

·       The author has continued to ignore the ICE/EV efficiency issue, which creates an error factor of about 4,

·       and has also taken the point that EdF are claiming an expected 25 TWh for Hinkly C, but simply made this one end of a wide range. As far as I can judge from correspondence, the 8 TWh figure seems to have been constructed in rather an odd way, by looking at the maximum output of any current UK nuclear power station, ie about 1.0 GW, and multiplying by 8000 as the approximate number of hours in a year. But of course Hinkly C is a planned 3.2 GW, and Sizewell B, which is the most recent of existing nuclear plant has a nameplate of 1.2 GW, ie about a third of the size. This has all the hallmarks of someone who does’nt understand the basic units of measurement, or indeed what they are doing at all.

·       The author has also invented a new figure of £ 50 billion as the upper end of the cost range for “a nuclear plant” – implicitly another Hinkly C. This implicitly acknowledges my comment that a £ 22 bn cost of Hinkly might be over the top anyway (and not just because of cost of capital issues), but only by putting in a lower end of £10 bn. The £ 50 bn is sufficiently large, when combined with top end of the range on stations required, to give a high final number that even exceeds the previous estimate of total cost. But both bits are absurd numbers which remain unexplained.

John Rhys.