As the leader of a first-term government that came to office in the biggest landslide in the state’s history, and with polling showing him to be, by some distance, the most popular leader in the land, New South Wales Premier Mike Baird would ordinarily be thought to be sitting pretty as campaigning for the March 28 election begins in earnest.

Instead, a pall of uncertainty surrounds the Baird government’s bid for re-election, notwithstanding the unbackable odds presently being offered by betting agencies.

For one thing, the comfort of a big majority was shown to be entirely illusory by the spectacular defeat of Baird’s counterpart in Queensland little more than a month ago.

For another, a spate of opinion polls timed to coincide with the start of the campaign period give cause to doubt whether Baird’s undoubted popularity will be enough to keep the inevitable swing back to Labor within manageable limits.

Polls from Newspoll, ReachTEL and Roy Morgan over the past week have sung from much the same song sheet in crediting the Coalition with between 43% and 44.6% of the primary vote, with Labor on 35% to 36% and the Greens on 10% to 11%.

A similarly solid consensus among pollsters on the eve of the Queensland election appeared to confirm the widely held assumption that the Liberal National Party would make it over the line, minus a very great deal of skin.

As was detailed here in depth at the time, the problem lay in the pollsters’ assumption that minor-party and independent preferences would behave as they did at the previous election in 2012. In the event, Labor’s share of preferences was so improved that its two-party vote was fully 3% higher than the polls had projected.

The million-dollar question now is whether the same phenomenon is set to play out in New South Wales. If not, Mike Baird’s only worry is the number of gold watches he’ll have to buy to thank defeated Coalition members for their four years of service, a repeat of the 2012 landslide being entirely out of the question. But otherwise, the two-party results from the most recent polls of around 53-47 in the Coalition’s favour should instead be read as 50-50 — in which case, it could very well be gold-watch time for Baird himself.

For historical perspective, the charts below track the estimated flow of minor-party and independent preferences at state elections in New South Wales and Queensland, the only two jurisdictions that grant voters the option of declining to preference one major party over the other.

 Following the introduction of optional preferential voting in New South Wales in 1981 and Queensland in 1992, most voters initially maintained their long-established habit of numbering every box. That came to an end with watershed elections at around the turn of the century, with the short-lived disturbance of One Nation providing a catalyst in both cases.

At the 1999 election in New South Wales, One Nation recommended its supporters number only one box in retaliation against the major parties’ directive to put them last. Since One Nation accounted for nearly half the non-major party vote on that occasion, this led to a dramatic upsurge in exhausted preferences.

In Queensland, it was Labor who provided the cue to voters to number only one box at the 2001 election, by way of distinguishing its own tightly unified operation from the mess of competing voices offered by the Nationals, the Liberals, One Nation and its short-lived spin-off, the City Country Alliance, which was formed after six of the 11 One Nation members who were elected in 1998 defected.

This appeared to have an educational effect with respect to optional preferential voting, causing the exhausted-preference rate to spike from around 30% to between 50% and 60% over the next five elections. Then came the unheralded reversal in January 2015, when a slump in the exhausted-preference rate was matched by a no less remarkable surge in preferences to Labor.

In contemplating whether we’re likely to see a repeat of that in New South Wales, it would seem that a case could be made either way.

It’s clear enough that the Abbott government will be hardly less of a weight in Mike Baird’s saddlebags than it was in Campbell Newman’s, its recovery in the polls over the past fortnight notwithstanding. One of the manifestations of this will surely be a stronger motivation on the part of Greens voters to maximise the damage they inflict on the Coalition.

It can also be anticipated that the highly successful union campaign which encouraged Queensland voters to “put the LNP last” will be pursued with equal or greater zeal in New South Wales.

On the other hand, the charts above demonstrate that preference behaviour in New South Wales has generally been less volatile than it has in Queensland.

It’s particularly striking that the extraordinary 2011 election result in New South Wales did not elicit any change in the behaviour of preferences, whereas Queensland Labor’s wipeout in 2012 was matched by both an unprecedented rate of exhausted preferences, and a record low preference share for Labor. This suggests there is less slack for Labor to pick up in New South Wales than there was in Queensland.

There is also reason to think an ambivalent attitude to the relative merits of the major parties will prove more persistent south of the border.

On the Labor side of the equation, the stench surrounding the party was always a lot sharper in New South Wales, and the work of the Independent Commission Against Corruption has made sure that names like Eddie Obeid and Joe Tripodi have remained in the forefront of the public mind.

Conversely, the O’Farrell-Baird government has not acquired the reputation for ideological severity that distinguished that of Campbell Newman (or indeed that of Tony Abbott), which inspired so many hitherto apathetic Greens voters in Queensland to deliver a second preference to Labor.