More econometrics on FuelWatch

The ACCC has updated their econometrics on FuelWatch and it reinforces the positive effects of that policy found in WA. The new updates do two things. First, they respond to claims that price falls in WA were the result of the entry of Coles rather than FuelWatch. The ACCC uses a test for endogenous structural breaks to look for significant events. They find that both Coles’ entry and FuelWatch were such events but that FuelWatch’s effect was almost three times as large. Now, readers of this blog — although you might not know it — are familiar with the use of these tests in competition policy. They were at the heart of my submission to the Reserve Bank on credit cards demonstrating that payment system reforms were not associated with structural break points even though normal market events such as interest rate rises were. (By the way, Stephen King and I have a new article on those reforms in the Melbourne Review).

Second, the ACCC looks at the highs and lows of the petrol price cycle. They find that the introduction of FuelWatch in WA, not only reduced prices by 3.5 cents per litre on high priced days it reduces prices by 0.7 cents per litre on low priced days. So that means that even bargain hunters were better off. This appears to be all about competition rather than about the petrol price cycle per se. All very interesting.

4 thoughts on “More econometrics on FuelWatch”

  1. I was interested to see that the price dropped on low price days, though it supported my prior that most of the effect would be on the high price days.

    It would be interesting, though, to see more descriptives on this (medians, modes, standard deviations)

    The data that would be really interesting is the quantity data – Zhongmin Wang’s (unpublished) work has some evidence on this

    A more public economics question remains is if Fuelwatch is worth the implementation costs (particularly given the second best issues that you’ve already raised).


  2. Freight markets spiked and have never returned to pre 2001 levels due to global shipping changes at the same time as the fuelwatch introduction in WA 1/1/2001. I suspect that the ACCC may be measuring this effect as the market discontinuity that they credit to fuel watch. The absence of freight from the ACCC analysis remains a serious flaw.

    this is the best link i could find on the web that shows this spike.

    Given that Perth’s ‘price cycle’ if you call it that is vastly different to the easterns states I am not sure how they compared the eastern price cycle with a very scrappy Perth market. Would be more credible if we had been given that actual analysis and not the conclusion since i remain unconvinced.


Comments are closed.