Errors in BP’s forecasts of the impact of electric vehicles


BP’s respected Energy Outlook was published this week (February 2018). Many commentators have written about the forecasts for electricity generation. I want to concentrate on the impact of electric cars and other efficiency gains on the demand for oil. Specifically, I’m going to focus on what I see as modelling errors and implausible assumptions in the BP analysis.


 Source: BP Energy Outlook, 2018, page 34

Source: BP Energy Outlook, 2018, page 34

Modelling errors

1)    BP sees electricity providing about 4.2% of all transport energy in 2040.[1] The figure is about 1.2% today, almost entirely arising from the electricity used in rail transport. So EVs (and possibly further electrification of rail and even air travel) will only add, BP says, 3.0% to electricity’s share of transport energy between now and 2040. (This small share is clearly demonstrated in the chart above). However, even using BP’s strikingly low figures for EV penetration, the company sees electricity providing the energy for about 31% of all car mileage and just under 15% of all truck travel by 2040.[2] This is a very striking – and arithmetically impossible – disparity.

Electric cars are, and will continue to be, more efficient at using energy to achieve movement than an internal combustion engine and I use a conventional assumption about the energy typically required to move an EV.[3] My calculations show that BP has underestimated the share of energy use contributed by electric cars by about 50%. Instead of rising from 1.2% of transport energy in 2016 to 4.2% in 2040, the true number (using BP’s assumptions) is about 7.2% of energy use.

In support of my assertion that BP has made an arithmetic mistake, I offer another comparison. The company says that natural gas will provide more energy for transport in 2040 than electricity. This is despite gas cars driving only 6% of the miles travelled by electric cars and 25% of the distances of electric trucks. Even though electric cars are probably about three times as energy efficient as LNG or CNG powered vehicles, this is wholly insufficient to explain what I think is BP’s error. (Please note that BP does state that shipping will also use natural gas even though virtually no ships are currently powered in this way).

2)    By contrast, BP has underestimated the efficiency gains from improved internal combustion engines. Once again, I am making this assertion only using the figures BP itself publishes in its extremely useful data file. In its calculations of the total demand for liquid fuels from internal combustion engine cars and trucks, BP assumes an approximate 34% efficiency gain across the entire parc of cars between 2015 and 2040.[4] Nevertheless, its own background figures actually provide an estimate of a more than 50% efficiency gain between 2016 and 2040. [5]The company also states that internal combustion engine efficiency improvements are speeding up. It suggests an average gain of ‘2-3%’per year between now and 2040, implying an approximate gain of around 45%.[6]

These differences are vitally important. BP forecasts a rise of 325 million tonnes of oil equivalent (MTOE) liquid fuel use in transport between 2015 and 2040, taking the figure from about 2,400 to about 2,700 million tonnes. If, instead of the published It correctly used its own estimates for efficiency gains for cars, this number would actually fall, even assuming lower efficiency gains in trucks. This change would, of course, adversely affect the global demand for crude oil. Instead of the picture BP presents of a rise in oil demand for transport to 2030 and then a very slow decline, volumes would fall much earlier to levels below today’s figures.

Highly questionable assumptions

3)    BP sees electric cars rising in number from about 7m in 2020 to about 95m in 2030.[7] This represents average sales of about 9m electric cars a year over the decade, assuming that almost all electric cars sold in the period are still on the road in 2030. This will be between 7 and 8% of all cars sold, assuming average global sales of around 120m a year during the decade. This is in sharp contrast to almost all other commentators.

UBS says, for example, that 16% of all car sales in 2025 will be electric.[8] Major German manufacturers have suggested that between 20% and 25% of all sales in 2025 be plug-ins.[9] Volkswagen talks of selling 3m electric cars a year by mid-decade, a third of BP’s estimate for global sales average across the decade. Several countries, recently joined by Ireland, are planning a ban on internal combustion engines by 2030.[10] Some states, such as Norway and China, may move even earlier.

4)    BP sees the average car travelling a rising number of miles each year. The almost two billion cars on the road in 2040 will each drive an average of about 16,200 kilometres a year, up from about 13,000 in 2016, a rise of over 20%.[11] This flies in the face of the long run downward trend in car mileages in developed countries. There is no justification provided for this assumption and it seems highly unlikely to come about. Doubling the number of cars on the road, and increasing the mileage each travels seems to conflict with the highly congested roads both in developed economies and urban portions of many newly industrialising states.


BP's mistakes and unconventional assumptions all tend to increase total oil demand for transportation in 2040 compared to 2016. These errors are all multiplicative. More plausible inputs and calculations would result in a forecast from BP that sees oil needs falling earlier than 2030 and then declining at a much faster rate than it projects from this peak. BP does also publish alternative scenarios for electric car sales but I argue that it should use more accurate numbers for its central forecast.


[1] Page 34 of the data pack at

[2] Page 38 of the data pack.

[3] I assume 6 km of travel per 1 kWh of battery electricity supplied. If we added a supplement to account for energy losses in recharging batteries the BP underestimate would be increased.

[4] BP uses 2015 figures, not from 2016 on page 34 of the data pack.

[5] Page 36 of the data pack.

[6] This is assuming a constant annual gain of 2.5%. The assertion that efficiency will rise by 2-3% a year is found on page 37 of the main presentation at

[7] Includes both plug-in hybrids and fully electric cars.




[11] Data from pages 36 and 38 of the BP data pack. I assume the car parc in 2016 was about 950 million vehicles.