A one-sided history serves as backdrop to the five federal by-elections on 28 July. Only once since Federation has the opposition party lost a seat to the governing party in a federal by-election. On that occasion, in Kalgoorlie in 1920, the Labor member had been expelled from parliament for uttering “seditious and disloyal” views about British policy and the Empire. He recontested the seat, and lost — a cruel and decisive blow to early twentieth-century republicanism.
In recent years the statistical picture has been complicated by the increasing tendency for Labor and the Coalition to sit out by-elections in each other’s safe seats. Where a genuine Labor–Coalition contest takes place, the size of the swing can depend on whether the seat is held by the government or by the opposition. Examining twenty-seven relevant by-elections in opposition-held seats since Federation, election-watcher Kevin Bonham finds a median swing of just over 1 per cent to the opposition, although that median embraces a very wide range of movement.
With the contests in Perth and Fremantle effectively Labor versus Greens, the results (likely Labor wins) are unlikely to tell us much about the state of the nation. In Mayo, recent SA election results might have augured poorly for Rebekha Sharkie (Centre Alliance, ex–Nick Xenophon Team) but local polling suggests that she is the favourite. Malcolm Turnbull has a little, but not a lot, riding on the result. A win would be nice, but a loss can easily be spun as the normal result for a government attempting to regain a seat in a by-election.
Braddon in Tasmania, held by 2.2 per cent, is the sort of seat that should favour a Labor opposition pursuing an equality and fairness agenda; it certainly doesn’t contain many voters consulting their accountants about the tax rate on $200,000 or more. The Liberals did extremely well here in the state election, but local experts point to a long history of contrasting state and federal voting behaviour, and the poker machine issue was also a live one in that poll. A Labor loss would be unwelcome, but could almost plausibly be filed under “Tasmania is different.” When Bob Hawke took government in 1983, he didn’t win a single seat on the island.
Most interest will focus on the Queensland seat of Longman, won by Labor in 2016 with the assistance of One Nation preferences, an anomaly not to be repeated this time. Hence, its margin of 0.8 of one per cent is probably artificial, and it may be that Labor itself needs a “real” swing to hold the seat. But it would be unwise to exaggerate the capacity of One Nation to control its preferences. In 2016, with the party card directing preferences away from Coalition incumbent Wyatt Roy, 56 per cent followed that advice but 44 per cent preferenced Roy regardless. Granted, more will be inclined to follow a pro-conservative card this time, but experience suggests that 70 per cent compliance is a likely maximum, which would still put the seat back in Coalition hands if nothing else changed.
It will possibly assist the preference flow to the Coalition if, as expected, more One Nation activists are available to distribute how-to-vote cards than would be the case in a general election, when resources must be spread across the vast state. If there is a donkey vote to be secured, it will be Labor’s, for Susan Lamb appears above the Coalition candidate on the ballot paper. (In Braddon, the Liberals are the donkey beneficiaries.)
Of the ten most marginal Coalition seats in the nation, five are in Queensland, and it is difficult to envisage a path to government for Labor that doesn’t involve increasing its stocks in that state. It might be contended that if Labor can’t prevail in Longman in the congenial anti-government atmosphere of a by-election (against a government that’s been trailing in national polls for most of its term), questions should be raised as to how it would achieve the necessary swing in the more demanding environment of a general election.
The state breakdown of Labor’s national poll lead is positive for the party. Indeed, the authoritative BludgerTrack identifies a movement to federal Labor of more than 5 per cent in Queensland since the 2016 election, a trend reinforced in the recent Fairfax/Ipsos poll reporting a two-party-preferred 52–48 in Labor’s favour in the state. Polling in Longman suggests a much closer contest, and the by-election will to some extent be a further test of the credibility of individual seat polling, following flawed predictions in the federal seat of Batman and the state seat of Darling Range in Western Australia.
Of course, it might be argued that Longman is different and that a general swing across the state may not be reflected in this particular seat: possible, but not obviously persuasive. There is some suggestion that voters may seek to punish Lamb for mishandling her original nomination and Bill Shorten for his hubris in declaring his members free of problems with their citizenship. If this factor is in play, Lamb may need to generate some level of sympathy for the strained family circumstances that led to her problematic nomination.
A more worrying possibility for Labor is not that Longman is different, but that Queensland is different, and that predicted levels of Labor support fail to materialise when the hypothetical of a polling question is replaced by the reality of a ballot paper. That has certainly been the case in previous elections, in which federal Labor has underperformed relative to opinion poll numbers. Only three times since 1949 has the party won a majority of the two-party-preferred vote in the state, and on only six occasions has it won a majority of seats. At present it holds just eight out of the state’s thirty House of Representatives seats.
Queensland has long been federal Labor’s boulevard of broken dreams. Bill Shorten will be hoping that a nightmare can be avoided. ●
