Sunday, June 01, 2008

FuelWatch: "The claim relied on by the Prime Minister is false"

The claim relied on by the Prime Minister that Western Australia’s FuelWatch scheme has cut the retail margin for petrol by an average of 1.9 cents per litre is false.

An examination by the Canberra Times of the methodology used by Australian Competition and Consumer Commission to arrive at that estimate finds that it hinges on an unusual definition of the word “average”.

In using the ACCC’s estimate Mr Rudd has often been careful to insert the words “up to” before “1.9 cents per litre” even though the ACCC’s documents themselves do not include the qualification.

Mr Rudd told Melbourne radio yesterday that the ACCC had examined seven years of Western Australia data and concluded that overall FuelWatch “represented something like a price difference between Perth and the eastern states of up to 1.9 cents a litre”.

The ACCC petrol price report released in December found that “the average of the price margin reduced by a statistically significant amount for Perth relative to the eastern capitals in the time since the introduction of FuelWatch”.

“The relevant weekly average price margin was around 1.9 cents per litre less on average,” it said.

The figure has been used to bolster the Government’s case against the Opposition and four of its own departments who have attacked plans to take the scheme national...

On Thursday the ACCC released updated modeling that further bolstered the government’s case concluding that that the introduction of FuelWatch to Western Australia in 2001 had cut prices by "an average of 3.5cpl for the highest price day of the week”, “an average of 0.7 cpl for the lowest price day of the week” and “an average of 1.8 cpl for the remaining middle five days of the week”.

The Prime Minister described that conclusion as “quite stark”.

But it hinges on an odd definition of the word “average”. The Canberra Times has confirmed that in preparing both sets of econometric analyses the ACCC calculated the “average” differential in price margins between Perth and the eastern capitals by adding up each price observed and dividing the total by the number of observations.

The method took no account of how often those prices were actually paid.

If, for instance, very little petrol was bought at the highest prices charged in Sydney and much more was bought at the lowest prices, the ACCC’s calculation of the “average” Sydney price would overstate the price actually paid in
Sydney and make Perth purchases look cheaper relative to eastern capitals than they actually were.

Indicative data on petrol purchases obtained by the Coalition’s Consumer spokesman Luke Hartsuyker suggests that Sydney petrol is in fact far more likely to be bought at times and places when it is cheap compared to Perth petrol.

During the month of August 2007 almost two thirds of the petrol sold in Sydney was sold for prices in the lowest 30 per cent observed.

In Perth only a little over one third was sold for prices in the lowest 30 per cent.

Expensive petrol prices were rarely paid in Sydney. There, only 3 per cent of petrol was bought at prices in the top ten per cent. In Perth almost 12 per cent was.

The ACCC did not have access to purchasing data for the full period of its surveys and so could not take purchasing into account in its calculation of “averages”, but the indicative data for August 2007 suggests that FuelWatch had a smaller effect on the average margins charged in Western Australia compared to the east than the 1.9 cents per litre quoted.

The Commission’s other finding – that the lowest margin charged in any given week fell by 0.9 cents per litre as a result of FuelWatch – remains valid.