1951 And A Reason To Avoid An Early Double Dissolution

In recent weeks there has been a lot of speculation about a possible early federal double dissolution, after the Greens and Coalition deferred Labor's Housing Australia Future Fund Bill 2023 until October.  The Albanese government says that this deferral constitutes a first "failure to pass" the Bill for the purposes of Section 57 of the Constitution and that if the Senate were to block or defer the Bill again, then the Bill would become a trigger for an early double dissolution election whenever the government wished to call one.  The Greens dispute that the deferral is a "failure to pass", but it would be a brave Governor-General who refused an election to a government that had an even arguable case that that box had been ticked, and in my view the government's case would be more than merely arguable.  There are questions about the mechanics regarding whether the Bill needs to be withdrawn before the Senate could fail to pass it again, but in some form or another it seems to me the HAFF Bill could become a double dissolution trigger if it is not passed in October or shortly afterwards.  

This article concerns not the mechanics of whether and how Labor can acquire an early double dissolution should it be unable to pass the HAFF Bill (though I am happy to have that discussed in comments), but whether it is a good idea strategically. 

 During the last term of the Morrison government it was generally obvious that a DD was a terrible idea for the Coalition as it would throw away the government's excellent 2019 half-Senate result and flood the Senate with even more minor party Senators than in 2016.  But for Labor, a DD now is considerably more thinkable in terms of the composition of the Senate.  If Labor could roughly repeat their 2022 result, then they would erase the 2019 slate in which they won only 11 state seats to the Coalition's 17.  The improved prospects for minor parties might not then mean that Labor's seat tally increased, but the DD would throw a bunch of Coalition seats to minor right parties and could also create extra avenues for (Labor+Greens+X) to pass Bills.  My estimate of a DD result if the 2022 vote shares were repeated is Coalition 28 Labor 26 Green 12 One Nation 5 Legalise Cannabis 2 UAP 1 JLN 1 Pocock 1.  Note that that is also only a few seats short of the level at which Labor plus the entire non-Greens crossbench is a majority, for what that's worth.

Enthusiasm for Labor improving its own seat position in 2022, with thoughts that the Greens should be very afraid as they could lose seats, has been seen since noises were made about the HAFF Bill as a DD trigger (example).  The argument is that on current polling Labor would be likely to gain seats.  However it is unlikely Labor could really maintain its current polling position or anything like it all the way to an early double dissolution.  The last three new governments that were dominating polling at the same stage in their careers (those of Hawke, Howard and Rudd - though Howard's polling honeymoon was just about to end at this point) all ended up with substantial seat share swings against them at their first re-election attempt, and 2PPs below 52%.  

In the case of Hawke 1984 - perhaps the most relevant comparison for a really early election of the sort being talked about now - there is a false belief that the poor result was caused by informal voting, and an at least exaggerated belief that it was caused by the length of the campaign.  It was most likely caused largely by the government simply being out-campaigned by an underdog Opposition, as witnessed by the remarkable turnaround in Andrew Peacock's personal ratings.  The last government to do really well - albeit with a swing against it - at an election at the end of its first term was Malcolm Fraser's in 1977. As of early 1977 (around the same stage of its life as the current government is now) Fraser's government was actually not polling all that well, with some polls even having the Whitlam Opposition ahead.

Australian federal elections also tend to be on average closer than their lead-up polling, and perhaps most importantly there is what I refer to as the Labor Fail Factor - the historic tendency of the Coalition, for whatever reasons, to on average do a much better job of recovering from bad polling to a good or at least not so  bad election result.  All up, even assuming the Albanese Government won the House of Reps election there is no guarantee it would even hold its Senate primary vote at the 2022 level, so while current polling holds out the possibility of gaining seats, history says it could even lose some (even on the assumption that they would be re-elected as first-term governments generally are.)

As for the Greens, if they could hold their vote at the 2022 level they could gain one seat at a double dissolution (the seat they lost when Lidia Thorpe defected). Gaining more would be very challenging, with a vote around 18% and a very favourable distribution of results for other parties needed to win three in any state (or around 20% without the favourable distribution).  If they regressed to 2016 or 2019 level support then they could easily lose two or three seats.  The HAFF impass might hold out the thought that the Greens' support could collapse even below that, but there's no evidence of that in current polling.  

I suggest the real reason a really early double dissolution (before mid-2024) is unlikely to be a great idea for the Albanese Government is mechanical, a reason that became apparent in a previous case of a government going to a DD during its first term (1951).  The terms of Senators elected at a DD are backdated to the previous mid-year point.  This means that for a DD held any time before mid-2024, the terms of the new Senators are backdated to mid-2023 and those terms expire in mid-2026.  By mid-2026 therefore, the Government would have to call either a third Reps election in four years or a standalone half-Senate election.  

While it might be argued that three Reps elections in four years would be bearable if the second was held very late year as an early DD (leaving a two and a half year gap to the government's second defence) that would require an election hot on the heels of or even entangled with the rancorous Voice referendum, an unappealing prospect especially if the Voice is defeated.  It's also doubtful whether the second attempt would have got far enough in the Senate process (which includes, eg, second reading speeches) for the box to be ticked in time for an election this year.  So the government would end up chopping about a year off each of its first two terms or having a half-Senate glorified by-election for voters to kick it in as-yet unknown economic waters - what does it gain that is worth this?

The argument I make here does not apply to a DD held in late 2024 or early 2025, which might turn out to be a good idea for some reason (maybe if there is some really crucial legislation that the Senate is for some reason blocking and that really needs a joint sitting to pass it).  But it would be unsurprising if by the time such dates appear on the radar, the government has stopped polling at honeymoon levels and the polls are no longer a source of such temptation.

1951 And All That ...

The Menzies Government created a similar timeline to the one I suggest Labor should avoid in its first term, but the comparison shows how different the stakes were.  Labor under Ben Chifley had gone to an election in December 1949 and lost, but retained control of the Senate in majority as a legacy of the undemocratic "slate" system used for the last time in 1946.  In the absence of a DD Labor's control of the Senate would stretch out to mid-1953 (three and a half years of the new government's life!) so Menzies' government used the referral of the Commonwealth Bank Bill to a committee at its second attempt before the Senate to immediately argue that the Bill had "failed to pass" and obtained a double dissolution for April 1951.  

The Menzies government won the double dissolution, obtaining a majority in both Houses and was able to pass the trigger Bill without needing a joint sitting.  But the double dissolution provisions backdated the new Senate terms to 1 July 1950, meaning that a half-Senate election would be needed by mid-1953.  

Enter the 1951 Fadden "horror budget".  Harsh medicine from it and a recession had such impact that the few polls available show the government falling into the abyss from late 1951 through the following year and well into 1953, including some of the worst polling a government has recorded in federal history. In February 1953 the government was down 39-60 in one Gallup poll (39.5% 2PP equivalent.)

A House of Reps election even in mid-1953 would have seemed unthinkable and the Government went to a half-Senate election instead.  Labor won 17 seats to the government's 15 but the government was carrying a 16-12 majority in the remaining long-term seats and clung on to a Senate majority of two.  Matters improved considerably for the Government later in the normal three-year Reps cycle and the polling suggests it was at least competitive and may have won anyway before the Petrov Affair came along on the eve of the 1954 election.  Then by 1955 Labor was wrecked by the DLP split and Menzies was able to call another early election to re-align Reps and Senate terms and win easily.

One might ask why Menzies did not, a la Malcolm Turnbull in 2016, hold the 1951 election in July so that his government would not have its future options so cramped by the impacts of the DD.  I don't know to what extent the economic circumstances of the time made this option strategically unwise, but Menzies' desire for an election was such that he would have hardly wanted to risk Labor changing its mind and deciding to pass the Bill after all.

Also a component of the advice to the Governor-General in such cases is to argue that the workability of Parliament requires a double dissolution.  This became clearer in 1983 when Malcolm Fraser initially omitted to advise the Governor-General sufficiently on the workability of the Parliament, delaying approval of his DD request while Labor was in the process of switching leaders to Hawke. (Contrary to some myth, the GG was satisfied that Fraser did have valid trigger bills.)  Obviously, the best time to argue that a blocked parliament has become urgently unworkable is the instance a DD trigger is available (which Menzies did).

In summary, through 1949-1955 the Menzies government bombarded the voters with elections and in the end was lucky or good enough to just get away with it, but this was triggered by a case where the Senate was obstructed by an Opposition majority.  It could be much harder for the Government of today to convince the public that sporadic holding out by the Greens on particular bills justifies cutting two terms short by a year. Alternatively a 2026 half-Senate election in these days of declining major party loyalty could be a gruesome spectacle.   Barring, for instance, a catastrophic split in the Liberal Party, it is hard to see how a very early DD for the Albanese government would be worth the risk in terms of restricting the government's electoral options down the track.

So even if the Greens do continue to play constitutional chicken with the HAFF Bill, I'm doubtful that the government will want to pull this trigger soon.  

--------------------------------------

NB It is irrelevant to this article's main purpose, but worth noting that an early double dissolution would result in a mini-redistribution in NSW, Victoria and WA (a normal election in August 2024 could also have the same result).  There was speculation in an op ed by Don Harwin that a possible impact of a mini-redistribution (subject to enrolment numbers) would be a temporary merged seat of Warringah-Wentworth which would pit teal independents Zali Steggall and Allegra Spender against each other.  I have checked the legislation and my view is Harwin is correct. A water boundary between these seats was created at the 2016 redistribution, prior to which the harbour was in the division of Sydney, and it appears that a water boundary is a boundary for mini-redistribution purposes.  (The AEC says it is and the Act does not say it is not!)  It is possible that the parliament could change the mini-redistribution rules before any early election.