The election result is in the mail

In Monday’s Crikey, Christian Kerr noted that while postal votes usually favour the Coalition at any election, this particularly seemed to be the case on November 24 last year. He wondered whether this was related to the ever-increasing benefits enjoyed by sitting MPs.

The first table below shows all declaration votes in seats Australia-wide. Dec votes are basically any votes that aren’t ordinary – they requires the voter signing a piece of paper.

Five seats – Calare, Kennedy, Mayo, Melbourne and New England – are excluded from this table. In these seats a full two party preferred break-down is not available either for 2004 or 2007 or both. (That’s why the total 2007 Labor two party preferred vote is given below as 52.9 when the correct nationwide figure is 52.7.)

Two party preferred votes by vote-type

What does this table tell us? The total swing to Labor was 5.6 percent (in reality it was 5.45 percent Australia-wide, but this lot has five seats missing, remember.) The subset of declaration votes that swung by the most amount was provisional at 7.4 percent. (We looked at provisional votes last month).

Postal votes swung by the smallest amount of any kind of vote – 5.1 percent. So postal votes did, in a relative sense, favour the Howard government. (They also of course favoured the Coalition in absolute sense, as they always do.)

But what about Christian’s incumbency theory? Below are swings in various categories of seats.

Two party preferred swings by vote-type and seat subset

There are several things to note about this table. One is that Coalition-held seats swung by more overall than Labor-held ones. This shows incumbency cannot be milked forever.

But the postal component of Coalition-held (before the election) seats swung by less than the overall amount, while in Labor-held seats the postal swing was a little more. This might indicate that government members had more success with postal votes than ALP members.

But note that Liberal inner urban (these tend to be archetypical blue ribbon, “leafy” Liberal) had a substantially higher postal vote swing to Labor than in ordinary votes. Probably the Liberals didn’t expend too much energy in majority of these which were safe.

But perhaps also, as postal votes are filled in days or weeks before polling day, this might indicate that these seats in particular swung back to the government in the last few days.

But the killer in seat outcomes is in the last column. These are electorates in which the Coalition just scraped in. While the national gap between postal and total vote swings was just half a percent, and across all Coalition-held seats it was also half a percent, in seats the Coalition won by less than one percent the gap was 1.3 percent.

This largely accounts for the Liberal Party prevailing in a bunch of several close seats such as McEwen, Bowman, Swan and Dickson.

We shouldn’t make big conclusions from aggregate data like this, and lots more fiddling is needed.

But it does seem that the Howard government’s postal vote strategy was very effective where it counted – in marginal seats.