South Australia lurches towards close election

A hung parliament is the result that the fewest voters want, but it's become the most likely in light of the lacklustre performances of the two major political parties in South Australia.

Analysis of our polling, conducted early in the campaign shows that while this election is superficially about jobs, the economy, and government's response to both, it is really more of a time for a change election, but this change may be smothered by doubts about the Liberal leader, and the overwhelming expectation that the Libs are going to win the election.

The graph below shows the top six issues, as nominated without prompting by our respondents. These are employment, the economy, government financial management, health, education and the federal government. Note that while Labor has been campaigning hard on trying to link the state liberals to Tony Abbott and the federal liberals, this is very much a minority concern, and restricted to those intending to vote Labor.

However issues aren't necessarily the reason people vote the way that they do. The table below gives the reasons people gave for how they were ultimately intending to preference – Liberal or Labor. This gives a much better idea of what is going on.

This graph demonstrates that Labor's biggest hurdle is the time for a change factor. Of those who responded who had yet to make up their mind, over 70% expressed an intention to protest, and in a state election protest votes generally go against incumbents. It was also a factor with Liberal voters. And while over the whole sample it was just less than 10% of respondents, it could be enough to determine the results in marginal seats.

Next strongest was the time for a change factor, dominated by Liberals. This is an argument that is almost conceded by Labor voters, whose strongest voting reason is the local candidate. This is followed by concerns about social justice and the Liberals.

A worrying factor for the Liberals is that being the "least worst" alternative is given more frequently as a reason for voting Liberal than Labor – their supporters are less enthusiastic.

The only two issues on this list that are specifically policy orientated is "business" – an intention to vote for the party that will be best for business – and "government" – concern about the size of debt and how the government is being run.

All of this suggests that the Liberals have not been successful in giving South Australians a strong reason to change. With the economy and jobs overwhelmingly the big issues, and the economy generally being a strong issue for the Liberals, they should really own this election, but they don't.

This may have something to do with the Liberal history in South Australia, where a number of respondents nominate factional problems as a reason for not voting for them.

While respondents acknowledge that Liberal leader Steven Marshall seems to have settled the factional problems down, he has his own problems because he is a first term MP. This means that voters really don't feel like they know him, unlike Labor's Jay Weatherill.

Weatherill's own Labor brand has been corroded, but he is seen as being a clean skin, unlike his predecessor Mike Rann. While some respondents nominate union links as a negative, most seem to feel as though they know Weatherill, and in a close election they will be weighing up the devil they know against the move for change of government.

Apart from their failure to put their policy mark on the election, the biggest risk to the Liberals is the overwhelming belief of electors that they will win.

The table below shows respondents expectations of who will win the election, weighted according to likely first preference voting intentions in the community at large (our respondents tend to lean heavily to the left, so we need to adjust for this), for the three largest parties.

While this table is not predictive of the actual result, it can predict voter behaviour, particularly in a situation where they have no strong loyalties one way or the other. Voters can reward a good member, even when representing a party they think should lose, because they believe that the other party will win, and so they can safely reward good personal service.

Couple this with the fact that just over 10% of respondents nominated the local member as a reason for giving their preference to Labor (and this was mostly to reward good performance, not to punish an under-performing Liberal member) and if the Liberals do not have a good marginal seats campaign, then they will fail to win seats that they would on a uniform swing.

With Newspoll showing the margin narrowing, and the South Australian electoral system having demonstrated around a 2% bias towards Labor in previous elections, this is going to make for a very close result. Indeed, it may be that Labor's marginal seat campaign is the reason that the boundaries seem biased – the Liberals may be winning big in safe seats, but failing to convert those on low margins.