Log in or register
My AA Account

Sign in to see your cover and request assistance online

Log in

Don’t have a My AA account?

You’ll need your policy or membership number

Create an account

Your Finances

Loans | Savings | Credit Cards

Your Driving Lessons

Book a lesson

fuel consumption

Can you rely on the official figures?

There's often disparity between manufacturer rates and reality

Recent studies have shown that the gap between 'real world' fuel consumption and that predicted by the official tests that car manufacturers must quote has been getting wider and wider over the past decade or so. The gap has increased from a relatively small 8% on average in 2001 to a remarkable 40% average in 2014. So the fuel consumption you can expect from your new car is getting on for half as good as you might expect if you take the official figures at face value.

No one standard test can accurately reflect the diversity of driving style, road, traffic and weather conditions, and drivers have probably always expected some variation compared to official figures, but in some cases the gap is now so wide that drivers are accusing manufacturers of misleading claims and even suggesting legal action.

It also appears to be the case that the more economical a car is to run, on paper, the greater the discrepancy between its theoretical and actual fuel consumption. This means that the official figures can’t even be relied on as an indicator of relative performance.

Fuel consumption

Type approval

Before a new car model can go on sale in Europe it is subject to a process called European type approval testing – tests to make sure the car, its sub-systems and components meet minimum safety and environmental performance standards. Test procedures are laid down in EU and UNECE regulations covering everything from whole vehicle crash safety to lighting and exhaust emissions.

One such test procedure determines the official fuel consumption and CO2 emissions data for the model. The car is driven on a rolling road under laboratory conditions, following a defined cycle of accelerations, gear changes, steady speeds, decelerations and idling.

Car manufacturers have to quote official figures in brochures and websites, and are not permitted to publish alternative figures. The official CO2 figure determines the VED (car tax) band for the model in this country and is used by the European Commission to track fleet average CO2 emissions and progress towards emissions reduction targets.

CO2 reduction targets

In Europe there are mandatory reduction targets for new car CO2 emissions. The fleet average must meet a target of 130g/km in 2015 (phased in from 2012) and 95g/km in 2020. Heavier cars are allowed higher emissions if they are offset by lighter vehicles with lower emissions - it is only the fleet average that is regulated.

From 2012, a manufacturer must pay a fine for each car registered if the average CO2 emissions for their fleet exceeds its limit value for that year. Manufacturers have a strong financial interest in reducing official CO2 emissions values as fines* for exceeding the target can be substantial.


Driving cycles

The current fuel consumption test follows something called the New European Driving Cycle (NEDC) which is in two parts, an ‘urban’ cycle and an ‘extra-urban’ cycle.

Urban cycle

From a cold start the car is driven following a rigidly defined stop start cycle covering a total of around 2.5 miles at an average of 12mph and briefly reaching a maximum of 31mph. The laboratory temperature is maintained at between 20 and 30C.


Following straight on from the urban cycle, the Extra-urban cycle also mixes acceleration, deceleration, steady speed and idling. The car briefly reaches a maximum speed of 75mph and covers a total of 4.3 miles at an average of 39mph.

Official combined fuel consumption
The test results in three official fuel consumption figures: urban, extra-urban and a ‘combined’ figure which is a weighted average of the other two.

Why aren’t the results accurate?

Many aspects of the test and test procedure contribute towards the growing gap between official figures and real world experience:

  • The basic driving cycle was first developed in the 1970s and, though it has evolved, it is short in duration and dominated by periods of idling, low acceleration and low engine load. It doesn’t adequately represent modern-day driving patterns and vehicle performance.
  • The test car only briefly reaches motorway speed so ‘highway’ driving is barely represented.
  • Air-conditioning, lights and other electrical loads – which increase fuel consumption in the real world - are all switched off for the test.
  • Flexibilities and tolerances in the test procedure can be exploited to achieve a lower overall fuel consumption figure – speed that is within the specified corridor for the test but consistently just below the target speed will result in lower fuel consumption than a speed that is just above.
  • The test is conducted in a laboratory at an ambient temperature between 20 and 30 C - considerably higher than the average temperature in the UK throughout the year.
  • The inertia class applied to the rolling road to simulate real-world vehicle inertia and aerodynamic drag, can be varied in discreet steps. A small reduction in vehicle weight, such as removing the standard spare wheel, may be enough to get the car just into the next lowest inertia class band and hence experience lower loads in the test than on the road. Research shows that a car is five times more likely to be just under an inertia limit than just over.
  • Car fuel consumption/performance can be optimised for the speeds/conditions of the test.
  • Test conditions may give a distorted view of the true benefits of new technologies - stop-start systems for example will show a relatively high benefit in a test in which idling is over-represented. The car is stationary for about 10% of the NEDC.
  • The test doesn’t take account of passengers or other loads
A new cycle and test procedure
Within the UNECE’s World forum for the harmonisation of vehicle regulations a new global driving cycle, the World Light Duty Test Cycle (WLTC) has been developed. The work has been lead by experts from Europe, India and Japan with the USA and China as observers.

New test procedures based on the WLTC should be in force and applied to new models from 2017.

The change in driving cycle alone (NEDC to WLTC) will result in an increase in fuel consumption/CO2 emissions of around 5% from the test but the effect of changes to test procedures and vehicle set-up, i.e. to reduce test flexibilities, will be more complex.

Performance in the current and new test will be compared but discussions are ongoing regarding how this will then be used to modify the Manufacturers' fleet average CO2 emissions as the target (95g/km in 2020) was set under the NEDC regime.
Real Driving Emissions (RDE)
European emissions regulations require emissions to be effectively limited under normal operating conditions and throughout the normal life of the vehicle rather than in type approval test laboratory test cycles alone.

To this end a new 'Real Driving Emissions' (RDE) test is to be introduced alongside the WLTP.

RDE uses a Portable Emissions Measurement System (PEMS) capable of measuring NOx, CO and particulate numbers on a vehicle while it is driven on normal roads in urban, rural and motorway conditions.

Test boundary conditions are still to be agreed as are so-called 'Conformity Factors' - how much higher should regulated emissions be permitted to be in an RDE test compared with in the laboratory cycle test?
Agreement on conformity factors
At the end of October 2015, the European Commission's Technical Committee of Motor Vehicles agreed implementation dates and conformity factors for RDE tests.

The new RDE test will apply to all newly approved types of vehicles from September 2017 and to all new vehicles from September 2019.

Conformity factors - the permitted difference between regulatory limits tested in the laboratory and values observed in RDE tests with real drivers on real roads - will be applied in two steps:

Step 1: a maximum 2.1 (110%) for new models by September 2017 (new vehicles by September 2019)

Step 2: a factor of 1.5 (50%) by January 2020 for all new models (new vehicles by January 2021)

In December 2012 the Director General of FIA region 1, the Director General of beuc and the Secretary General ANEC wrote to the vice president of the European Commission expressing concern that the current type approvals procedure using the NEDC doesn’t represent realistic driving conditions.

Read the letter in full (pdf)»

The way in which official fuel consumption is measured will change but this will only apply to new models - car models already on sale wouldn’t be re-tested.
Buying a new car

If you can’t trust the official figures that car manufacturers quote, where do you go to get an idea of the fuel economy you can expect to achieve with a new car?

Performance relative to official figures will vary between models and manufacturers but you could start by applying a simple rule-of-thumb correction.

If you assume that a new car will be capable of fuel economy around 40% less (worse) than the official combined figure quoted by the manufacturer you shouldn’t be disappointed, and may be pleasantly surprised.

And of you're looking at a used car the gap will depend on the age of the vehicle - around 10% for a 10 year old car up to around 25% for a car that's 3-5 years old.

No one standard test can accurately reflect the diversity of driving style, road, traffic and weather conditions, and no single driver is representative of every other – you and your neighbour, both owning the same car, would probably see different average fuel consumption figures.

But if you could see what tens or even hundreds of drivers are achieving in the real world in the same model, then the range, best to worst, and average would give you a fair idea of what realistically can be achieved, if you take your own driving style into account.

There are two websites that can help you here, one English, and the other German.

Real MPG (Honest John) – users have uploaded more than 33,000 fuel consumption figures. On average cars achieve 87% of the official combined fuel consumption figure but this includes models from the 1980s, 90s and early 2000s when the test was more representative.

Spritmonitor.de – slightly different in that users upload fuel purchase and mileage data and the site then calculates fuel consumption.

The site is reasonably easy to navigate once you apply an online translator but, being continental, all fuel consumption figures are quoted in litres/100km rather than miles per gallon.

To convert litres/100km to mile/gallon

Divide 62.15 by the fuel consumption in l/100km and multiply the result by 4.546

For example

5l/100km = (62.15/5)x4.546 = 56.5mpg
4l/100km = (62.15/4)x4.546 = 70.63mpg


(updated 5 January 2016) 

*The fine per car registered amounts to €5 for the first g/km of exceedance, €15 for the second g/km, €25 for the third g/km, and €95 for each subsequent g/km. From 2019, the cost will be €95 from the first gram of exceedance onwards.