During federal parliament’s 2016–19 term we lived through no fewer than nine by-elections, all but two of them thanks to section 44 of the Constitution. That’s the largest number in any term since Federation with one exception: ten were held in 1951–54, a remarkable nine of them due to deaths, including those of former prime ministers Billy Hughes and Ben Chifley.
The two-party-preferred swings in the 2016–19 term ranged from 7.2 per cent towards the government — specifically towards Barnaby Joyce, who was Nationals leader, in New England — to 7.0 per cent towards Labor after Malcolm Turnbull’s resignation in Wentworth. But the Wentworth figure was counted out by the electoral commission purely for interest’s sake: the Labor candidate didn’t make it into the final round, and it was independent Kerryn Phelps who defeated Dave Sharma 51.2 to 48.8 per cent.
(Because Phelps didn’t run at the previous election, there could be no two-candidate-preferred swing either. Yes, one figure was widely reported — the difference between Malcolm Turnbull’s 2016 two-candidate-preferred vote and Sharma’s in 2018 — and at 19 per cent it made great headlines, but it was meaningless.)
In Bennelong the previous December, Labor’s Kristina Keneally had achieved a 4.8 per cent swing, which wasn’t too bad considering sitting Liberal John Alexander had recontested. Little was made of the result. Scribes didn’t furiously plot that 4.8 per cent on the pendulum to produce an oh-my-God massive Labor victory at the next general election. But they did after Longman, Queensland, eight months later, despite the swing to the opposition being a smaller 3.7 per cent (assisted in this case by a new personal vote for Labor MP Susan Lamb, who had been elected in 2016). Longman set the hares running — apply that 3.7 per cent to seats across the board, and woaaah! — and resulted in prime minister Malcolm Turnbull’s demise two months later.
But the truth is that by-elections can rarely tell us anything about a party’s prospects at the next general election, precisely because voters aren’t (except in rare cases, in a hung parliament) deciding who should form government. This frees them up to vote for other reasons.
And folks should never ever plot a by-election swing onto the pendulum. Yes, it’s tempting, but don’t do it! Yet virtually everyone in politics is convinced, or pretends to believe, that by-elections are dry runs for the big event. Perhaps we got this unfortunate habit from the Brits.
This takes us to Eden-Monaro, where voters will cast their ballots on 4 July to select Mike Kelly’s replacement. Five weeks ago I ventured that, “factoring in the likelihood of a strong [Liberal] candidate,” the government would probably retain the seat.
That strong candidate (an Andrew Constance or a Jim Molan; I didn’t mention state Nationals leader John Barilaro, but he would have fitted the bill) didn’t eventuate, and instead they’re running Fiona Kotvojs, who also contested in 2019. (Not that she’s a bad candidate, but she doesn’t possess star power.)
Labor, by contrast, seems to have wrung maximum bang out of its buck with the mayor of Bega Valley Shire, Kristy McBain. She is reported to have enjoyed a profile boost during the summer bushfires (though not as much as Constance), and around a fifth of Eden-Monaro’s voters live in her jurisdiction. (And for all I know her recognition might extend well beyond the council boundaries.)
The key theme of that earlier article was that Kelly’s popularity, and hence his high personal vote, made the seat appear more Labor-friendly than it actually is. Redistributions over the last decade and more have favoured the Liberals, and Kelly’s presence on the ballot paper at all of the last five outings (even in 2013, when he was unsuccessful) disguised this reality.
Put it this way: if Kelly had retired at the last election, Kotvojs (if she had been candidate) would now be the MP. If he’d waited and retired at the 2022 general poll, the Liberals would probably retake it, unless Labor has a particularly good New South Wales result (like 2007, say, or 1993).
But this is a by-election, and by-elections are unpredictable, though they usually swing against governments. Even if we take away 5 per cent for the loss of Kelly’s personal vote, just a Bennelong-sized swing would get Labor over the line.
So, in the absence of that turbocharged candidate, I’m withdrawing the earlier prediction and climbing onto the fence. It could go either way.
From McBain’s point of view, the chances of entering federal parliament next month are better than if Kelly had delayed retiring until 2022. And if she does win, it will increase her chances at the next election because she’ll have almost two years to consolidate her personal vote and employ the resources of incumbency. Defending a seat is easier than trying to take it.
What issue could Labor urge Eden-Monaro voters to “send a message” to the government about on 4 July? At time of writing, industrial relations potentially holds promise.
Needless to say, Anthony Albanese has more to lose from Eden-Monaro than Scott Morrison, because seats usually swing to oppositions at by-elections, don’t they?
A Liberal win wouldn’t tell us anything about the next general election, but lots of political players will believe it does, and others, Albo’s internal enemies, will pretend to. •
And if you want to dig in a bit deeper, see this graph of votes at groups of Eden-Monaro booths over the past sixteen years.