FuelWatch: evidence-based-policy or policy-based-evidence?

Economic Papers (Economic Society of Australia), Dec, 2008 by Don Harding

1 Introduction

Stilgoe, Irwin and Jones (2006) provide a detailed discussion of evidence based policy in the United Kingdom and the United States that sets the context for this paper. They document cases where expert evidence was often reshaped and repackaged by governments so that it supported existing policy rather than informing policy decisions.

The decision to introduce a national FuelWatch scheme raises a number of issues about the interplay between evidence and policy in Australia; see Harding (2008a, 2008b, 2008c).

In this paper I focus on just two questions. First, was the policy shaped by the econometric evidence or was the Government's presentation of the evidence shaped by the pre-determined policy? Second, is the econometric evidence sufficiently robust as to support the FuelWatch policy?

2 Background to FuelWatch and the 2007 ACCC Report on Petrol Pricing

The national FuelWatch scheme which was announced on 15 April and is to start on 15 December 2008 is based on the Western Australian Government's scheme of the same name that was introduced on 2 January 2001.

2.1 The 2007 ACCC Report on Petrol Pricing

Petrol pricing has been a contentious issue in Australia and has been the subject of a large number of inquiries. In many of these inquiries the ACCC has voiced suspicion that the Western Australian FuelWatch scheme might lead to increases in the average price of petrol. In late 2007 the ACCC delivered a report on Petrol pricing that was very cautious about FuelWatch stating that:

  Assessing any system in the style of FuelWatch that incorporates
  increased price information and price commitment requires great
  care due to the potential for anti-competitive as well as
  pro-competitive benefits. Although the inquiry gained a preliminary
  assessment of the impacts in Perth from the scheme, it is clear that
  a case-by-case approach is required to assess the potential impacts
  on competition of any similar scheme. In particular the ACCC has not
  analysed the application of such a scheme to rural and regional
  areas. Apparent extra considerations here include the increased
  potential for anti-competitive effects due to the more concentrated
  nature of the market, the extra cost in initialisation,
  administration and compliance and how to decide which areas to cover.
  In summary, there are potential benefits and potential costs of
  adopting a national price commitment arrangement that need to be
  carefully considered. (2)

The powers that were conferred upon us by then federal Treasurer, Peter Costello, powers of subpoena, powers of ability to obtain evidence from Informed Sources that had never been available to us before and powers to obtain data that enabled us to undertake econometric modelling.

Appendix S of the ACCC report provides an econometric analysis of the Western Australian FuelWatch scheme. Changes to the relevant laws in 2006 and at the end of 2007 provided the ACCC with new powers to subpoena documents, data and witnesses. (3) Mr Graeme Samuel, Chairman of the ACCC, explained the purpose of that econometric work to the Senate Estimate inquiry as follows:

  There has been an enormous focus on one particular half-page of
  our report, the econometric analysis and appendix S of that report.
  The econometric analysis which was described in appendix S was
  undertaken with one purpose in mind, and that was to determine
  whether or not FuelWatch had caused any harm to Perth motorists
  since its introduction. It was designed to assess whether FuelWatch
  had caused any increase in prices to Perth motorists (Senate
  Committee Hansard, Standing Committee on Economics, Estimates,
  Thursday 5 June 2008).

The ACCC claim that their analysis is robust because it has been subject to scrutiny within the ACCC and by Treasury. (4) However, neither the Senate Estimates hearing nor the subsequent Senate Economics Committee hearing could obtain a comprehensive description of the actions taken by Treasury to scrutinize the ACCC econometrics.

In many fields of endeavour information is not viewed as being evidence unless the results can be replicated independently--what this means in practice varies but a minimum requirement is that the data be public and the paper must have been subject to independent public scrutiny. Neither of these minimum requirements were met in the case of FuelWatch.

Experience with 'evidence based policy' making in the United Kingdom (UK), see Stilgoe, Irwin and Jones (2006), confirms that the absence of public scrutiny often led to government agencies filtering out information that was inconsistent with government policy. The UK experience also suggests that internal review procedures and even review by other government agencies have little value in protecting against such 'policy-based evidence'.

2.2 Government Decision to Introduce FuelWatch

On 15 April the Australian Cabinet decided to implement a national FuelWatch scheme. The press release announcing that decision said that (5)

  Econometric analysis undertaken by the ACCC last year concluded that
  under the WA FuelWatch scheme the relevant weekly average price
  margin was around 1.9 cpl [cents per litre] less on average.