It’s been a long-standing article of faith at this site that Queensland state politics will be dominated by Labor until the Liberals elbow the Nationals aside and assume their rightful place as the senior coalition partner. But given the Nationals’ use of their institutional dominance to defend the status quo, it was hard to see how this was supposed to happen. For this and other reasons, the news that the two parties have been engaged in two weeks of secret merger negotiations has come as a profound shock.

The Nationals’ seniority in the Queensland Coalition is a legacy of circumstances that have ceased to apply: the state’s traditionally decentralised population, the rural malapportionment that was abolished when Labor came to power in 1989, and the personality cult of Joh Bjelke-Petersen. The first of these factors is disappearing due to the prolonged boom in the urban south-east, where the population has grown 65 per cent in the past 20 years, from 1.7 million to 2.8 million. This is double the rate of growth in the rest of the state, where the population has risen from 900,000 to 1.2 million in the same period. Just as significantly, the growth in the south-east has been largely driven by interstate migration, which has drawn in voters who have no historical affinity with the Nationals. These newcomers have erased the memory of Bjelke-Petersen’s great political successes: his incursion into suburban Brisbane at the 1983 and 1986 elections, and his maintenance of the National/Country Party stranglehold on the Gold Coast despite the area’s post-war urbanisation.

As a result, the urban branch of conservative politics is becoming more important to the Coalition’s electoral prospects with every passing year. But this has not been reflected in the parties’ representation in parliament, where the Nationals have maintained the greater numbers throughout the electoral convulsions of the post-Fitzgerald era. In large part, this is the result of a vicious cycle in which the Liberals suffer electorally because they are seen as subordinate to their country partners, who have the advantage of a support base in areas impervious to challenge from Labor. This has deprived the Liberals of bargaining power in the important negotations to determine which seats are contested by which party. Such agreements are necessary because Queensland’s system of optional preferential voting does not compel voters to direct preferences, making three-cornered contests lethal for the Coalition. These agreements continue to freeze the Liberals out of important seats in Brisbane and on the Gold Coast, despite their overwhelming dominance there at the federal level.

The attraction of a merger is that it resolves these problems without demanding a surrender from the Nationals, whose state parliamentary leader will remain as Leader of the Opposition (though Graham Young at National Forum notes that the move will “structurally mean the dissolution of the Queensland National Party, with its assets and members being transferred to the Liberal Party”). But the effective disappearance of the Nationals also raises serious electoral problems that may yet queer the deal. Not for the first time in Australian political history, it became fashionable a few months back to talk of the Nationals’ impending demise following Victorian Senator Julian McGauran’s defection to the Liberals. I wasn’t persuaded then and I’m still not now. The urban/rural divide is the most important cleavage in Australian electoral politics and probably always will be, owing to Australia’s unique concentration of people and power in a small number of state capitals. Country voters have never been willing to suffer representatives they perceive as subordinate to the dominant city interests, and they are not about to start doing so now for the sake of Coalition unity. Their desire for a distinct voice will continue to find expression in one way or another, and the Liberal Party would be better off having it harnessed by a coalition partner than surrendering it to external forces.

I don’t think Peter Beattie meant to be helpful in saying so, but he hit the nail on the head with his assessment (as quoted in the Courier-Mail) that a merger “would spark the re-emergence of One Nation-style parties and independents”, who would exploit the perception that the Nationals had “sold out the bush”. Not surprisingly, this point is well understood by the Prime Minister, who the Courier-Mail reports is “yet to be convinced about the merits of a merger and may seek to oppose it”. If he does so, things could get very messy very quickly – so much so that Graham Young raises the possibility of Peter Beattie cashing in on the turmoil with a snap election.

UPDATE: Graham Young reports that the “New Queensland Liberal Party” “appears to be still-born”.

Piers Akerman of the Daily Telegraph has thrown the cat among the pigeons by reporting that "political minds with close ties to the Howard camp" are talking of an "elegant departure" by the Prime Minister at the end of the year. Akerman, who for various reasons is known for the quality of his Coalition sources, paints an impressively detailed scenario in which Howard moves to the back-bench in December to assume an "elder statesman" role and avoid a by-election for his seat of Bennelong. Peter Costello and his new Treasurer (most likely either Brendan Nelson or Alexander Downer) would thus be given "time to work together" ahead of an early election to be held "in late March or April, before the next budget".

Firm talk of an impending Howard departure is new, but the early election aspect was echoed a fortnight ago in Crikey, which said "talk in some Liberal circles says a February federal election should not be ruled out". This comes as a surprise because, as Akerman in particular should well know, an election before the first week of August 2007 is all but out of the question. This is because the six-year terms of the Senators elected in 2001 (who took their seats in mid-2002) will not expire until the middle of 2008, and the election to replace them cannot be called until the final year of the term. Due to the minimum period required for an election campaign, the earliest possible date for a normal election for the House and half the Senate is August 4.

It is techically possible for a House-only election to be held before a half-Senate election is due, but the only time a government has willingly done so was in 1963. Bob Menzies was then surviving on a one-seat majority after his government’s brush with death at the 1961 election, so he could credibly claim he was seeking a fresh mandate when he moved to take advantage of Labor’s internal ructions over state aid and its indecisive response to the establishment of the US base at North West Cape (which culminated in the "36 faceless men" episode). The 1963 election succeeded in restoring the Coalition to a comfortable majority, but it put the two houses out of alignment and required separate mid-term half-Senate elections to be held until the clock was reset by the 1974 double dissolution.

Not surprisingly, the Coalition performed poorly at the mid-term elections, which loomed as "national by-elections" of a type that any government would prefer to avoid. After the second such election in November 1967, the Coalition was reduced to 27 seats in the 60-seat chamber, having earlier held between 30 and 32 in the years since the 1951 double dissolution. The November 1970 election weakend its position further, leaving the Coalition with 26 seats and the Democratic Labor Party with five. It would be very odd behaviour for a government with a handsome majority to burden itself with such difficulties for the sake of getting an election in before the budget, especially if (as present indications suggest) the budget loomed as another revenue-gorged bonanza of tax cuts and giveaways.

The other scenario for an early election, a double dissolution, is even less attractive. The Coalition’s once-in-a-lifetime Senate majority would instantly disappear, and they would need to poll around 50 per cent in each state at the ensuing election to maintain the strength of Senate representation to which they have become accustomed. Furthermore, the government would have to indulge in unseemly contrivances to meet the requirement for a double dissolution, namely the Senate having twice rejected a law passed by the House. It is true that company law amendments have been blocked with help from Nationals Senator Barnaby Joyce, and that a trigger could be created if another unsuccessful attempt was made to pass them. But would Joyce be willing to block it again if it led to his six-year term being cut short less than a third of the way through?

Given the practical difficulties, any talk of an early election emanating from the Liberal camp can only be a tactical ploy to keep Labor off balance and foment its leadership tensions. With that in mind, one aspect of Akerman’s article can be readily dismissed. This raises the question of whether the rest of it can be as well.

Sunday 4.00pm. The table has been amended to factor in late results and new intelligence on preference flows (thanks Kevin), and also to correct a calculation error that was inflating Labor’s vote after preferences.

7.38pm. Not sure if they do notional preference counts on election night - I could forgive them if they didn’t. In any case, we’re unlikely to get much more action tonight.

7.33pm. Creek Road, the last booth other than Launceston, has strengthened Parkinson still further in what looks pretty much like a status quo result.

7.29pm. All booths in from Rowallan, Hall is still on 82.3 per cent. Only waiting on Creek Road and the Launceston booth for primary votes in Wellington. Still no two-party preferred count.

7.25pm. A bunch of larger booths plus pre-polls are now in from Wellington. The primary vote swing against Labor has eased a little, such that I think we can safely call it for Doug Parkinson.

7.15pm. 74.0 per cent counted in Rowallan, Hall down ever so slightly to 82.3 per cent.

7.11pm. Preferences amended. I’ve now got 50 per cent of independents’ preferences going to the Greens and 30 per cent to Labor, the other way around for CDP, with the remainder exhausting.

7.10pm. A big flurry of booths in from Wellington, but the existing score hasn’t changed much. Labor are down about 5 per cent and the Greens are more or less steady. My arbitrary estimates showed Labor doing better on preferences than last time. I think I might amend that.

7.02pm. 60.8 per cent counted in Rowallan - possibly the quickest count I have seen, indicative of how this district is dominated by small towns. Little change in the figures: 82.5 per cent to 17.5 per cent.

6.54pm. 54.4 per cent now counted in Rowallan, Hall on 82.2 per cent, Greens on 17.8 per cent.

6.51pm. Six booths in for Wellington. My 5.0 per cent swing to Labor is based on pretty arbitrary preference calculations that probably flatter Labor. For the time being it might be safest just to trust the primary figures, which suggest Labor are losing votes to independents.

6.39pm. Oh dear - a sudden blurt of booth results for Rowallan after a quiet period. Hope it’s not like that for Wellington, of which there is still no word. 22.4 per cent counted in Rowallan, Hall leads 82.1 per cent to 17.9 per cent.

6.23pm. Those small town booths in northern Tasmania are coming in at a rapid clip. Seven booths now in, Hall leading Cassidy 685 to 130.

6.22pm. There are actually three booths in already from Rowallan, which is pretty impressive. Greg Hall leads the Greens 166 votes to 47.

6.20pm. All eyes are on Tasmania this evening as the events that have captured the nation’s imagination over the past two weeks build towards an emotional climax. That’s right folks, it’s the annual periodic elections for the Tasmanian Legislative Council. Since the seat of Rowallan looms as a certain victory for independent Greg Hall (his sole opponent is the Greens’ Karen Cassidy), the focus of attention here will be on Wellington, where it is at least theoretically possible that Greens candidate Marrette Corby or independent Hobart City Council alderman Marti Zucco could pull off an upset. Hopefully the booth arrangements haven’t changed too much since the last poll in 2000, and the calculations in place for the above table will prove of some use. Otherwise I will get rid of it and focus on off-the-cuff commentary. The Greens need a swing of 9.5 per cent to win the seat; if Zucco does better than expected and outpolls them, my table will not be of much use. First figures should be in around 6.40pm.

The Jeff Kennett will-o’-the-wisp came and went before I had the opportunity to comment on it. Long-time Kennett antagonist Stephen Mayne surveyed the damage in today’s Crikey email:

Late this morning, after less than a day of frenzied speculation, Jeff Kennett formally withdrew from the Victorian Liberal Party leadership race. Which is the worst possible outcome for the state opposition and leaves them almost certain to go further backwards at the 25 November election. Rather than someone like Ted Baillieu emerging as the consensus great white hope after Robert Doyle’s resignation yesterday, the electorate now knows that most Liberal powerbrokers believe he was a worse alternative than recycling a controversial premier. For the Labor Party of course, he’s a much better alternative - to fend off Jeff, they might have needed to dig into their cash pit with a well-resourced scare campaign. With Baillieu, the Bracks spin machine will hardly need to get out of second gear, let along go into overdraft, to retain office. Imagine the scenario if Jeff had come out yesterday and immediately ruled out a comeback on the basis that Ted Baillieu would make an outstanding Premier. Instead, we had all this frenzied expectation - and now nothing more than deflation.

Crikey also underlined the overwhelming consensus that Kennett’s return would have done little if anything to avert another Coalition disaster at the coming election (to say nothing of the absurdity of the proposal that he lead the party in the meantime from outside parliament). Allow me to add my voice to the throng. The common Liberal complaint that the 1999 election result was a "protest vote that went too far" is revealing more for its arrogance than its insight. The theory should have been laid to rest four weeks later by the Frankston East supplementary election, held because the sitting Liberal member died on the eve of polling day. Voters on the day knew perfectly well that a "protest vote" would sign the death warrant of the Kennett government, but they nevertheless delivered the seat to Labor with a swing of more than 7 per cent.

Kennett’s approach to the election campaign suggested that he saw it as an opportunity to build his Melbourne-centric personality cult, and to rub his enemies’ noses in what he saw as a looming triumph. This manifested itself in a number of ways – in the "Jeff’s a fuckin’ legend" pitch at the demographic of Formula One and Triple M, the latter of which was given regular access to the Premier while the ABC was snubbed; in the energy directed at winning the normally safe Labor seats in Dandenong that had been made temporarily marginal by the 1992 and 1996 elections, while the Coalition’s own marginals were neglected; and worst of all, in Kennett’s petulant performance on Jon Faine’s ABC Radio program three days before the election. Kennett presumably imagined that this would only be heard by un-Victorian basket-weaving leftists, but the footage that appeared on that night’s television news bulletins did incalculable damage to his image, particularly in the country. All the while Labor was making hay with its devastating advertisements on country television depicting two taps, one dripping slowly and marked "country Victoria", the other gushing freely and marked "Melbourne". It is unlikely that country Victoria has forgotten what it perceived to be its neglect at the hands of a Kennett government fixated on bread and circuses in the capital. In failing to recognise this, Kennett’s boosters are showing the same short-sightedness that proved so costly in 1999.

There is no objective reason why the result of the 1999 election should have come as such a shock. The late opinion polls were mostly on the money, with Newspoll and Morgan correctly indicating a dead-heat and only ACNielsen erring in favour of the Coalition. The sense of surprise can be put down to the Melbourne media’s assumption that the election would be won and lost in the traditional battlefield of Melbourne’s eastern suburbs. Instead, swings in the country of up to 10 per cent delivered wins to Labor that few had seen coming, including Gisborne, Ripon, Seymour, Narracan and Ballarat East. The following table indicates geographic variations in the swing to Labor at the 1999 and 2002 elections using the newly created regions for the upper house. The methodology for the calculations was a bit slapdash, but the results are useful for illustrative purposes.

It can clearly be seen that the country gave the Coalition its worst results in 1999 and its best results (relatively speaking) in 2002. This has left a fair bit of low-hanging fruit for them in the country, specifically Evelyn, Hastings and Gembrook (all held by margins of less than 2 per cent) (UPDATE: and all arguably outer urban, as noted by commenter Geoff R) along with Morwell (4.9 per cent) and South Barwon (5.0 per cent). The recovery of these seats is essential to any kind of respectable performance, and would most likely be jeopardised in the event of a Kennett comeback. If the Coalition is to go further and actually put the Bracks government in jeopardy, there are a further eight country seats it must win that are held by margins of between 6.8 per cent and 9.5 per cent, and these would surely be beyond Kennett’s powers.

On the other hand, Kennett might have strengthened the Liberals’ position in 11 Melbourne seats with margins of between 2.1 per cent and 5.8 per cent, all of which are located east or south-east of the city. If Kennett had inspired a swing of 5 per cent to 6 per cent in these areas that was not reciprocated elsewhere, he could have added a respectability to the scoreboard out of proportion with the overall statewide swing. Such a result would have reflected the outcome of the 1996 election, when Labor failed to yield dividends from a 2.8 per cent swing due to another poor performance in the eastern suburbs. This concealed the Coalition’s weakened position and contributed to an exaggerated perception of Kennett’s electoral record.

An unrelated point on the Liberal leadership: earlier this week I received an email from an ABC reporter in Melbourne looking to pick my brain regarding Tuesday’s Newspoll, which showed Steve Bracks leading Robert Doyle 60 per cent to 15 per cent as preferred Premier. In particular, he wanted to know what became of other leaders who had polled this badly. One encouraging precedent for Doyle came to mind, namely Queensland Nationals Leader Rob Borbidge. Going into the 1995 election, Morgan had Labor Premier Wayne Goss leading Borbidge 70 per cent to 17 per cent (in February), 72 per cent to 15 per cent (April) and 74 per cent to 16 per cent (June). Then came the election on July 15, at which the Coalition outpolled Labor 53.3 per cent to 46.7 per cent on two-party preferred. Six months and one by-election later, Borbidge was Premier. The most widely credited factor in this surprise outcome? The Goss government’s insistence on proceeding with a hugely unpopular toll road. Perhaps Doyle should have hung in there after all.

In other news, the elections for the Tasmanian Legislative Council districts of Rowallan and Wellington will be held tomorrow, though neither is likely to be of much interest unless the Greens or Hobart alderman Marti Zucco can pull a rabbit out of the hat in Wellington. This site will provide some sort of live coverage, although it remains to be seen whether my trade-mark results tables and swing calculations will prove feasible.