February 13, 2004

Nationals – forward to the past

I asked the question at the end of my last post why the Liberal Party allowed the National Party to contest seats on the Gold Coast when it is not naturally National Party territory. The answer goes to the heart of the coalition problems. The Liberal Party hierarchy was not tough enough in negotiations. It also had factional reasons for not running in Broadwater. And the National Party is overly ambitious.

I wasn’t in the room when the deal was done, but presumably it went along similar lines to our negotiations with the Nationals in 1994 and 1995. They would have first staked a claim to the seats for sentimental reasons – “We held them last”. The correct response to that is “So what. You’re only as good as the last election and you lost them. Let’s do some research to see who would win.”

From the public record Caltabiano appears to have played that card. Then there would have been argument over the research. Research on these matters isn’t accurate without candidates being named, but neither side would have wanted to reveal their potential candidates because the other side would have put pressure on the prospective candidates either not to run, or to switch sides. It happened last time. So the research issue goes nowhere.

Then there would have been the argument that the National Party wouldn’t be able to hold back the local branches and they would endorse candidates anyway. In fact, the threat would be that they would end up endorsing candidates all over the country causing a slew of three-cornered contests. With optional preferential voting, they would have said, that would mean that you would lose all these seats.

This is generally called blackmail, and the way to deal with it is to treat it with contempt or counter-threat. It is a lose/lose proposition, so is one that only an irrational negotiator would seriously advance as more than a gambit. You therefore call their bluff. Mind you, it is exactly the proposition underlying the Nationals “Just Vote 1” strategy which was aimed at bullying people into voting for them rather than independents.

It is the key National Party negotiating tactic in Coalition negotiations. So central is it that when the Liberals tried to change the system back to the more democratic compulsory preferential during the Borbidge government they were opposed by the Nationals, even though the Nationals will tell you that optional preferential is a bad system.

I believe that the National Party would have backed down if the Liberal Party had stood their ground. To make certain of this the Liberals probably needed to threaten to field candidates in the Toowomba North, Lockyer, and all the Townsville and Cairns seats. They could have given up all of these with no damage to their position because they were seats less likely to be won than the Gold Coast ones.

They would have had even less trouble if they had constructed a realistic narrative of what each party should be doing which left them with the Gold Coast (and Sunshine Coast) seats. That narrative should have been that this last election was about each party getting back to its roots and dealing with its specific challenges. For the Nationals this is taking on the Independents and One Nation, and for the Liberals it is winning back the urban heartland. In this division of Labor the Gold Coast falls to the Liberals, as it should.

So, given my division of labour, and given the obvious failure of the Liberal Party in its half, how did the National Party fare? They performed reasonably well. In rural areas their vote was up and they now hold 9 of their 15 seats by margins of more than 60% two-party preferred. Some of the swings in these safe seats were 10% or more, reflecting in part the decline of One Nation. Hinchinbrook is a good example of this.

Independents were contained in the “sugar seats”, although as I have suggested above, by a blackmail strategy. This may have bludgeoned independent voters in the short-term, but it is not a strategy to win their hearts and minds. Certainly where there were viable Independents, such as in Nanango, Nicklin, Gladstone and Gympie, the swing was against them. One of the two surviving One Nation members, Rosa Lee Long was also too much of a challenge, increasing her margin.

In urban areas the results were mixed. Kurwongbah and Kallangur on Brisbane’s north saw reasonable swings to them off a very low base. In Logan and Springwood on Brisbane’s southside they barely received a swing at all. In most of the other urban seats in southern Brisbane they also had only small movements to them, or small movements away. This was similar to the Liberal Party’s experience in Brisbane, but the Libs clearly outperformed them on the Gold Coast.

Now, having laid out most of the groundwork, tomorrow I’ll finish with where the Coalition should go from here. As the challenges at the next election will be very similar to the challenges at the last election this prospective analysis will also really be a retrospective one.

There’s not much point giving advice to Beattie. On his results he doesn’t really seem to need it!

Posted by Graham at 12:07 AM | Comments (4)

February 11, 2004

One Nation and the Liberals – going, going…?

Participants in our focus groups were almost uniformly dismissive of One Nation, so we predicted that they would not do well outside particular seats. They did better than we thought. Their average result for the seats that they ran in was 8.7%. This is a little deceptive as Bill Flynn with 20.76% and Rosa Lee Long with 47.58% raise the average considerably, so it’s worth looking at the median which is 7.68%.

To put it into perspective this average is actually slightly higher than the Greens average of 7.9% and higher than the Greens median of 7.19%. One Nation also won a seat, while the Greens didn’t even come close. Why is there a widespread perception that they did worse than the Greens? Presumably it is a measure of momentum as well as their lower profile without Pauline Hanson. Last election the Greens scored a statewide vote of 2.51%, so support is growing. One Nation scored 8.69% across the whole state, which implies a higher average vote per electorate than this time, so their support is shrinking.

Even with momentum running against them this is still a fairly impressive figure. However, Independents had an average of 10.1% in the seats where they ran, meaning that the average Independent candidate did better than a One Nation one, and of course five of these Independents actually won a seat. In fact the results of those who won distort the average position of independents. The top five Independents averaged 52.07% of the vote in the electorates where they won, and the bottom five only averaged 0.72%. The median Independent vote was 3.52%. So running as a One Nation candidate guaranteed most a better vote than they would have got running as an Independent.

One Nation’s vote was also found fairly eclectically around the state, once again demonstrating the fact that they are a phenomenon in both urban and rural, Labor and non-Labor areas. They peaked in Tablelands and Lockyer, with Callide and then Kallangur (on Brisbane’s outskirts) next. Their worst result was in Indooroopilly.

The Liberal Party scored an average of 34.15% in the seats where it ran. This went from a low of 13.2% in Inala to 52.93% in Moggill. This in itself tells a story. The ALP won 42 seats with an absolute majority, Independents 3, the National Party 9, but the Liberal Party could only manage 1. This is the quantitative measure of the contempt that showed in some of our qualitative research. Absolute majorities are a good proxy for how strongly you hold your heartland. While the Liberal Party had a similar average per seat to the National Party, their voters don’t appear to love them the way that National Voters love the National party.

Another interesting statistic is that out of the 21 seats where the Liberal Party received 40% or more of the two party preferred vote, only 4 of them are in Brisbane. Another 4 are on the Gold Coast, and 2 on the Sunshine Coast. It has been apparent for a while that it has ceased to be a Brisbane Party. Liberal Party State President Michael Caltabiano has been suggesting that if the Liberal Party had been allowed to run in all of the Gold Coast and Sunshine Coast seats that the Coalition would have won more of them. This is an ultimately untestable hypothesis, but there are a few straws in the wind.

The only two seats on the Gold Coast where there was a swing to the government were Burleigh and Broadwater. The National Party contested both these for the Coalition. In the seat of Southport, also contested by the National Party, the there was a 1.67% swing away from the Government. Currumbin has to be put to one side as a special case. In the other seats that the Liberal Party contested they received swings to them of 5.25% in Mudgeeraba, 4.2% in Robina (Leader Bob Quinn’s seat), and 1.88% in Gaven. This supports the hypothesis and suggests that Broadwater and Burleigh could have been Liberal wins as they had margins of 2.5% and 1.9% respectively.

That the Gold Coast is natural territory for the Liberal Party is one that I have argued for quite some time. It is a view that I share with the Gold Coast Bulletin. Twelve months ago this paper was writing editorials telling the National Party that it was not welcome on the Gold Coast anymore. This begs the question. Why did Caltabiano agree to them running there in the first place? Something to start with tomorrow as a prelude to looking at the Nationals’ vote.

To be continued…

Posted by Graham at 04:17 PM | Comments (1)

February 10, 2004

Independents and Greens

Our research suggested that there was much more volatility in rural and regional Queensland than in the southeast corner. This, with some notable exceptions like Currumbin, was true. Our research also led us to believe that Independents in these areas would poll well but win little, while Greens in the southeast corner would poll well and win nothing. We didn’t believe that antagonism to the government would go in any substantial way to the Opposition in either area. This was not as generally true as the first two propositions.

Sugar seats

On Saturday night at the tallyroom there was a confrontation between Bob Katter and Ron Boswell. We had a new category of seat this year called “sugar seat”. I suspect that it got its name from Bob Katter, because everyone referred to “his” independents as the “sugar” independents. It’s pretty flexibly defined and no-one really seems to be clear exactly who is and isn’t a sugar independent. If anyone has a precise list, please email it to David Fraser and me and we will put it to good use, but in the meantime these paragraphs are based on my sense of what could properly be called a “sugar” seat.

The argument from Bob was that his string of candidates had performed well, given everyone else a lesson and as a result the sugar industry was going to be rescued. Boswell interjected that Katter had cost the National Party seats which meant he had achieved nothing. Who is right? The answer appears to be neither really. With the exception of Andrew Lancini in Hinchinbrook and Jeff Knuth in Burdekin, none of the independents in seats dependent on sugar did particularly well. And in both Hinchinbrook and Burdekin the National Party won.

Just Vote 1

Why did the Nats do so well when our polling showed that voters in these areas were not happy with the Opposition? It might be that we didn’t understand our polling properly, always a possibility. It might also be an effect of the “Just Vote 1” strategy. Last election this strategy worked in Labor’s favour and disadvantaged the Coalition, One Nation and Independents. The non-Labor parties divided the anti-government sentiment between them, but because preferences didn’t flow they split what might have been a majority into a collection of minorities.

This election the Coalition took a calculated risk. It told its supporter to “Just Vote 1”. Their strategy was then to leverage their likely size as the largest non-Labor grouping and argue that if you voted for an Independent you would just waste your vote. Quite possibly Boswell’s interjection was a deliberate reinforcement of that strategy. That doesn’t mean that voters responded to the mandate and enthusiastically voted National, I think they did it grudgingly. This might be one reason why the votes of the six established independents (with the exception of Lex Bell who lost) and Rosa-Lee Long, an independent-like member of One Nation, increased substantially. In areas where a popular Independent was likely to have the largest non-Labor vote, they received a huge local endorsement as voters abandoned both the Coalition and Labor would-be electoral duopolists. Liz Cunningham won 54.93% of the first preference vote, beating the Nats down to 7.3%. Peter Wellington did even better with 59.73%. We’ll never know how well Chris Foley did because his 65.29% first preference vote was partly due to the fact that the National Party disendorsed their candidate for his seat.

As an aside (I’ll follow up on this tomorrow) the National Party also used the optional preferential system to blackmail the Liberal Party into not ceding them a number of seats and avoiding three-cornered contests altogether.


As predicted the Greens did do well in a couple of seats, but not well enough to win. It would have been difficult for them to win because of the Liberal Party’s “Just Vote 1” strategy and they probably needed to achieve around 40% in a seat to have a chance of winning without a flow of preferences.

They should have been able to get close to that figure in Mount Coot-tha or South Brisbane judged on the New South Wales results, but failed. I would tend to put that down to the lack of a high profile respected leader, newspaper reports of internal wrangling, and no effective on-the-ground campaigns. We did get signs of this in our focus groups. While Drew Hutton is the most high profile Green, he does not command respect the same way that someone like Bob Brown does, and participants will tell you this.

There were attempts from sections of the Liberal Party to negotiate preference deals with the Greens in key seats, with a suggestion for a while that the Liberal Party would not run in South Brisbane. If the Liberal Party had left South Brisbane to the Greens there is a possibility that Anna Bligh could have been pushed hard with a decent campaign. Bligh only polled 53.41% on primaries, not good in such a safe seat. It would have been a real coup for the Liberal Party if the apparent leadership succession in the ALP could have been disrupted, not that Labor appears to need to think about a succession until sometime after at least the next election.

Tomorrow I’ll have a look at One Nation and the Liberals…

Posted by Graham at 12:00 AM

February 09, 2004

Beattie has a huge result, particularly in marginals

These are brief preliminary notes on the state election result and are the first of a series. We will keep the blog running for a little while yet as we analyse the results and look at some of the fallout. It will be some weeks before we can actually quantify the results of the election in terms of overall two-party preferred vote. At the moment we can only make best guesses on the results. This is the first of a series of shorter articles that will progressively examine the issues in the campaign. David Fraser will update his analysis at a later date, including the pendulum, this piece and following ones are all my own work. In this first piece I want to start to set the scene.

Before I do, some of this is going to be a bit technical, so here are the conclusions for the "time poor":

  • The two-party preferred swing is in the vicinity of 3.6% against the government

  • Beattie will have somewhere around 56% of the two-party preferred vote

  • There will be a net loss to the Government of three seats and a net gain to the Opposition of 5

  • The swing to the Coalition was mostly outside the marginal seats

  • More marginal seats under 3.6% moved away from the Coalition than moved to them

  • The Coalition marginal seats campaign was a disaster on the basis of its results

  • Swings

    To analyse the result we first need to work out what happened. The various party spinners, and a number of the “psephologists” are making this difficult because of the way they are using the word “swing”.

    Naturally everyone is trying to make their result look as good as they can. So the ALP is claiming that there was only a 1.74% swing against Beattie. This is based on their first preference vote. The National Party are claiming a 2.53% swing to them and the Liberal Party a 4.2% swing. These are all largely irrelevant figures on their own because they ignore the effect of minor party preferences. To a large extent the Coalition swings are really only the effect of the decline of One Nation and demise of the City Country Alliance and return of the anti-Labor vote previously trapped in there. They don’t represent the same degree of success in taking votes from the Government.

    My best guess is that the two-party preferred swing against the Government, allowing for minor party preferences and exhaustion of votes, will be in the vicinity of 3.6%. This is a huge victory for Peter Beattie. Originally he was sitting on a two-party preferred vote of approximately 60%, and after this election it will still be 56%. No other state government has achieved this sized majoriyt in successive election in the last 20 years.

    Position of the Parties

    Another way of looking at the election result is to look at the wins and losses of the parties, which is where it is really devastating for the Coalition. At this stage, and I wouldn’t expect anything much to change in any but perhaps Charters Towers, the Coalition has won 6 extra seats - Currumbin, Charters Towers, Surfers Paradise, Burdekin, Burnett, and Lockyer while the Government has won one – Keppel. This is a net gain to the Coalition of 5, but a net loss to the Government of only three because Lockyer was held by One Nation, and Surfers Paradise by an Independent. The Fraser Pendulum shows Lockyer on the Labor side of the divide on a notional distribution of preferences, but the seat has never been held by Labor, and I suspect that our pendulum’s notional distribution of preferences overstates the real notional margin in Labor’s favour. In another of those seats – Currumbin – it is probably more accurate to say that it was lost by the government, rather than won by the Opposition, given the furore surrounding Merri Rose before and during the campaign.

    One way of putting the swing into perspective and to judge the Opposition campaign is to look at what the results of a uniform 3.6% swing against the government should have been. What this shows is that the campaign was a failure in the seats where it should have been targeted, and that the swings actually occurred in the safer Labor seats. According to our pendulum, a swing of this magnitude should have delivered 13 seats.

    Instead the coalition won only three in this group. Of the others, Noosa, Burleigh, Toowomba North, and Broadwater moved away from the Opposition. In Nicklin, Gympie and Tablelands the Independents strengthened their positions. Clayfield moved around 1.5% to the Opposition and Indooroopilly a similar amount.

    Springborg claimed in his concession speech that the Opposition is now much closer to winning government. This is correct in direction but not magnitude and is only marginally encouraging for them. On our assessment they needed a uniform swing of 12.3% to win government this election with Hervey Bay as the swing seat. Next election, they could win minority government on a uniform swing of 7.61% with Toowoomba North the pivotal one. Yet their results in marginal seat campaigning suggest that they are as likely to do this as to put a man on Mars. There has been some shrinkage in the range of margins of safety on the Government’s side across all seats, but this election most of the marginal seats under 4% actually moved further away from the Opposition. That points to a failed campaign.

    To be continued…

    Posted by Graham at 12:01 AM | Comments (5)
    Graham Young
    John Black
    Mark Bahnisch
    Michael Lee