Email from Malcolm Mackerras, 15 June 2009

My reason to write is to tell you of my personal disappointment in relation to the Western Australian state election, held on September 6 last year. Notwithstanding the considerable amount of work I have put into analysis of this event I have not been able to get anything published since the election.

You have published a table of mine titled “Per Cent Share of Two-Party Preferred Votes for Defeated Governments – 1975 to 2008” which you described as a “great table”. Today I ask you to publish that table again, together with the two WA pendulums I sent to you a while ago. Those pendulums were done by the graphics department at “The Australian” newspaper.

The last time I had an article published on Western Australia in “The Australian” newspaper was as long ago as May 17-18 of 2008 in “The Weekend Australian”, the “Inquirer” section on page 27. The article was titled “One vote, one value” and dealt with the redistribution of seats. That article carried the first pendulum.

The essential reason for my disappointment is that the WA Electoral Commission did not count out the two-party preferred votes in the seats of Alfred Cove, Blackwood-Stirling, Central Wheatbelt, Eyre, Kalgoorlie, Kwinana, Moore, Nedlands and Wagin until all uncertainty was removed from the close result. Eventually they did those counts and we can now say that the swing against Labor was 4.2 per cent, that is, 52.3 per cent (see first pendulum) reduced to 48.1 per cent. See post-election pendulum.

The delay in the counts was a disappointment to me. However, there was another disappointment. In my letter of request to them I asked the WA Electoral Commission that, in the two-party preferred counts, those for Wagin, Central Wheatbelt and Blackwood-Stirling should be done on the basis of National versus Labor. For some strange reason, however, they defied all logic and conducted those three counts on the basis of Liberal versus Labor.

Doing that had two effects. The first is that the official aggregate figure for Labor is 48.1 per cent where (I contend) it would have been an even 48 per cent if those counts had been done sensibly. That also affects the predictive power of the pre-election pendulum. If the swing had been 4.3 per cent then the pre-election pendulum would have predicted the result exactly. By that I mean the deviations would have cancelled out exactly.

The second effect of the Commission’s perversity is to create wholly fictitious swings to Labor in Wagin, Central Wheatbelt and Blackwood-Stirling. If you compare the two pendulums you could calculate that there were swings to Labor in nine seats. Since three of those are fictitious I say that, in reality, there were swings to Labor in six seats, of which Albany is the truly impressive case. The six seats are Albany, Armadale, Mandurah, Mindarie, North West and Willagee.

The effect of the Commission’s delays was to make it more difficult to persuade “The Australian” to publish an article seen to be so stale. However, their graphics department did do the pendulum for me so I can see no reason why you should not publish it – given that they are clearly not going to do so.

Following the recent Fremantle by-election I had hoped they might be interested – but clearly they are not. One of the things which annoyed me was to read on two occasions their Perth-based journalists write that “Fremantle is the only Australian lower house seat held by the Greens”. That, of course, is wrong. There are four Greens in the Tasmanian House of Assembly and there are four in the ACT Legislative Assembly.

In my re-drafted article submitted to “The Australian” (clearly not going to be published) I wrote as follows: “Fremantle is now the only single-member constituency held by the Greens, either in Australia or New Zealand. All other Greens parliamentarians, both in Australia and in New Zealand, owe their seats to some form of proportional representation electoral system.”

Since the people who matter at “The Australian” assure me that it is their policy to publish my pendulums I assume they will publish this WA pendulum some time late in 2012 or early in 2013. It would, of course, require at least one change. Fremantle would be shown as held by the Greens (though its placement would not change, being just below Nollamara and just above Girrawheen) and the number of Labor members would be shown as reduced to 27.

Assuming no further by-election changes the left hand (government) side would read Liberal 24, National 4, Independent 3, Total 31, Majority 3 - and the right hand (opposition) side would read Labor 27, Greens 1, Total 28.

Kind regards

Malcolm Mackerras