George Monbiot's Guardian article on the motor industry undermines a fair point by citing 'facts' which range from the misleading to the downright wrong.

Monbiot's article (also available, with source references, on his own website) decries loans made by the US government to US car manufacturers to help them develop more environmentally friendly vehicles, and similar loans proposed in Europe.

He says the motor industry has, in the past, deliberately sabotaged new technologies and has missed targets for CO2 output and fuel consumption.

And these are points that are worth debating.

But Monbiot then undermines his own argument with preposterous comparisons and dubious 'facts'. Let's take, for instance, his assertion that 'The average car sold in the States today is less efficient than the 1908 Model T Ford'.

Monbiot's claim is based on a report of a Detroit News story which claims the Model T could achieve 25mpg and data from the US EPA which says the average consumption of 2008 model year 'light duty vehicles' (ie cars and light trucks) in the US is 20.8mpg.

A reasonable comparison? Not even close.

1911 Model T ad: claims 20-25 miles per US gallon (Image: Ford)

For a start, we need to examine in what way fuel consumption is a measure of 'efficiency'. Which of these is 'more efficient':

• A bus, carrying 50 people, which consumes fuel at 7.3 mpg (based on US bus data, but converted to Imperial gallons)
• A car, carrying four people, which consumes fuel at 20.8 mpg (to use the EPA's estimate for US cars)

The answer to the question, of course, depends on how you define efficiency. If we're talking about fuel use per passenger per mile, the bus wins by a wide margin - even though it has poorer fuel consumption than the car.

If we came up with a definition of effiency which included not just fuel used but also the time saved by the people travelling, the bus would probably be less efficient: even Monbiot agrees in his own article on badly-organized bus travel that a journey by car can be a more efficient use of time.

Monbiot bases his 'efficiency' argument purely on fuel consumption, but the comparison is only valid if the two vehicles have similar capabilities, and that is quite simply not the case. The Model T has, by modern standards, poor weather equipment, poor brakes, poor roadholding, non-existent secondary safety, laughably short service intervals and a hopeless lack of reliability - not to mention minimal performance.

Give any major car manufacturer's engineers the chance to build a car to the same specification and it would easily trounce the Model T for fuel consumption and it would generate far fewer harmful emissions.

Then we have to consider whether the two figures - 20.8mpg for modern cars, 25mpg for the Model T, both referring to US gallons - have been arrived at using comparable methods. And there's no evidence that they have.

The modern car figure is a 'real world' estimate by the US Environmental Protection Agency. There's no reference to tell us where the data for the Model T came from or how it was derived. How far can we trust either of these figures?

You might think the EPA numbers would be rock-solid, but even the EPA itself casts doubt on them. The document Monbiot cites rubbishes the numbers on page 1 (my emphasis):

The projected fleetwide average MY2008 light-duty vehicle fuel economy is 20.8 miles per gallon (mpg)... based on pre-model year sales projections provided by automakers.

More so than in any other recent report, EPA believes that the pre-model year 2008 sales projections ... do not accurately reflect the actual light-duty vehicle market in MY2008. Automakers submitted MY2008 sales projections to EPA in the spring and summer of 2007 when average nationwide gasoline prices were in the $2.50 to $3.00 per gallon range. Actual gasoline prices have averaged about $3.50 per gallon during MY2008, or $0.50 to $1.00 per gallon higher than at the time automakers provided sales projections to EPA. Based on publicly available sales data, which are not part of the formal EPA database, it appears that higher gasoline prices have led to a 10 to 15 percent decrease in overall light-duty vehicle sales relative to automaker projections. Further, the sales data suggest that subcompact, compact, and midsize cars have been the only vehicle classes to have met or exceeded sales projections by automakers, while sales of midsize SUVs, large SUVs, and large pickup trucks are 15 to 25 percent lower than automaker projections. It also appears that 4-cylinder engines have gained market share from 6-cylinder and 8-cylinder engines. Accordingly, it is extremely likely that the projected fleetwide average MY2008 fuel economy value of 20.8 mpg is too low.

So the EPA reckons the sales estimates used to derive the fuel consumption figures are probably wrong, because market conditions have changed since some of the assumptions were made.

Did Monbiot not read the report properly, or did he simply decide not to reveal to his readers that it casts doubt on its own figures?

So, can we trust the Model T consumption figure? Ford claimed in adverts of the time that the Model T was good for '20 to 25 miles per gallon'. I'd guess a Model T would return about 25 miles per US gallon trundling along the road - but that's not the same as an EPA 'real world' estimated figure. Trundle along the road in a modern car - even a gas-guzzling V8 - and you'd get the same, or better.

The Model T wouldn't even get through a modern EPA consumption-measuring driving cycle - because the cycle average speed is 48mph, and a Model T is all-out at 45...

It's unlikely, then, that the figures have been produced using comparable methods. So you can't sensibly compare them.

So the result is this: we have figures which are not reliable, derived using methods which are not the same, for vehicles which are not comparable.

Now that, to use a phrase Monbiot likes, is junk science.

Vehicle taxation

Then there's Monbiot's assertion that 'cars were taxed at £1 per horsepower: in real terms (and in some cases in nominal terms) a far higher rate for gas guzzlers than today’s'. The article notes go on to explain that 'The top standard rate of vehicle excise duty from 2010 will be £455. The Mercedes-Benz SL is 604hp; the Lamborghini Murcielago is 640.'

Monbiot cites a report from the UK government's Environmental Audit Committee on vehicle taxation as his source for the comment that cars were taxed on horsepower. Indeed the report says 'The tax disc was introduced in 1920, and the tax charged at a graduated rate of £1 per horsepower.'

That's true, up to a point. In fact, cars were rated using the 'RAC horsepower' system: the RAC horspower is calculated from the bore of the engine (d) and number of cylinders (n) using the formula:

Because this formula made a number of assumptions about engine performance, it quickly became outdated as engine development progressed. Vauxhall's famous sporting car of the 1930s, the '30/98' had an RAC rating of 30hp but actually produced 98hp.

It's misleading, then, to suggest that cars were taxed on horsepower (unless you go on to qualify exactly what you mean by horsepower). Yet that is exactly what Monbiot has done in his article.

It's also misleading to say 'the Mercedes-Benz SL is 604hp': there are five SL models, and four of them produce less than that.

Monbiot has some interesting things to say. But he does his arguments no favours by using dubious data - and by talking about areas of engineering and automotive history he clearly doesn't understand.

An edited version of this article appeared in the Guardian's Response column on 17 October 2008.