A disturbing letter
The issue of whether the Howard government is, as the Labor party has claimed, the highest-taxing government in Australian history is not only politically contentious, but statistically complicated. A couple of weeks ago, David Bassanese had a piece in the Fin using Australian Bureau Statistics data to argue in support of this claim.
I was disturbed today, to see a letter from Rob Edwards of the ABS, responding to Bassanese and supporting the government line. In particular, the letter accused Bassanese of unspecified errors and argued in favor of relying on cash measures of the deficit (Bassanese used accruals for recent years and “cash converted to accruals” estimates for earlier years). While the ABS has occasionally responded to direct criticism of its figures (for example, my own criticisms of its multifactor productivity estimates), I don’t recall a previous instance where it’s been involved in partisan controversy of this kind.
As readers of this blog will know, the issue is far too complex for simple answers like those put forward by Bassanese and Edwards to be regarded as definitive. But Bassanese is a journalist writing what’s clearly intended as commentary rather than news. He’s entitled to put forward his own views. Edwards is supposed to be a neutral public servant.
It’s been a long time since I took on trust anything coming out of policy departments like Treasury and the Productivity Commission. Under the present government, we’ve already learned we can’t trust statements from the armed forces, the Defence Department or the Electoral Commission. But until now, I’ve never seen any serious evidence of political interference with ABS. This letter suggests that the process has begun.
(The Edwards letter follows).
I refer to David Bassanese’s “Howard writes tax history” (April 2), which used data provided by the Australian Bureau of Statistics.
The data provided to Bassanese for some periods was incorrect, and was a mixture of accruals and cash-based data. Accrual data are available only from 1998-99, whereas cash data are available for all periods. For analysis of government finance statistics covering long periods we recommend that cash data only be used, to maximise the consistency of the data.
Using cash data for all periods, and with corrections to the data in error, commonwealth tax revenue, as a proportion of GDP, in the period before the introduction of the GST was highest in 1986-87, at 24.4 per cent, and not in 1999-2000 as stated in the article. Using the revised data, the proportion for 1999-2000 is 24.1 per cent.
Australian Bureau of Statistics,
fn1. My own view, which I’ve stated a few times is, “if it isn’t, it should be”.