A good start

There are a number of reasons for that.

One is that expectations ran too strongly in Abbott's favour. As the pre-poll votes show, a swing similar to the Queensland or NSW state elections was on, but disappeared in the last days.

If voters think a result is going to be too strong one way or the other they will vote strategically to narrow the gap. One sign of that is large swings in safe seats, and small or no swings in marginals – there is plenty of evidence of that happening.

Another is that Abbott ran a "Rudd-lite" strategy, just as Rudd ran a "Howard-lite" strategy in 2007.

While this has the benefit of making it difficult for people to find reasons not to vote for you that aren't reasons not to vote for your opponent, the small target strategy also opens up options for third parties prepared to take a risk.

Many voters have a perception that politics is a confidence trick run by two identical teams who just swap the premiership between themselves. This made micro parties, particularly Clive Palmer's Palmer United Party, attractive.

That allowed voters to vote against Labor, but then direct a preference back to them. In the absence of Palmer more than 50% of them would probably have voted Coalition.

The small target strategy also makes it difficult to claim a mandate for too many things.

Another factor affecting the vote, particularly in Queensland, has to be the state government. While there is little evidence that Campbell Newman has cut anything severely, the unions have been waging a relentless PR campaign alleging this to be the case.

The perception that Newman did this, while promising not to in the campaign, and the reality that Labor ran an advertising campaign alleging Abbott would do the same, has to have had some impact, and could easily account for the slender margins of safety for Labor in seats like Moreton and Lilley.

Incumbency also played a factor.

So, how should these results play into government?