Insane In The Balmain: Does Labor Really Need To Win This Seat?

 

A Labor flier from Balmain (source)

The district of Balmain (Green 10.0% vs ALP) would normally be of little interest at this NSW election given that it is on a large margin and the Greens have so far never lost a single seat electorate that they won at a previous general election.  What makes Labor think Balmain is worth taking a tilt at is the retirement of 12-year incumbent Jamie Parker, the first Green at state or federal level to retire from a single-seat electorate.  

One of the reasons the Greens have 14 wins from 14 attempts at defending state or federal single-member seats won at a general election (by-elections and multi-member seats are different stories) is that there tend to be oversize swings to them in subsequent elections.  For a first defence at state and federal levels the average swing is 2.9% on primary vote and 4.0% on two candidate preferred.  For a second and later defences it is 3.0% and 4.8% per election.  (The 2CP figures are affected in varying directions by Liberal preference recommendation changes, and a 2CP swing is not always available).  The obvious explanation for this is demographics, but another possible explanation is that sitting Greens may receive unusually high personal votes.  And if they receive them when they defend their seats, maybe they also lose them when they retire?  

(Upper House results could be used to get an idea of whether this really is a personal vote effect as opposed to a demographic one.  I haven't tried, the point for this article being that the seat has attracted interest because of the retirement.)

Labor have been targeting Balmain with signs reading "Vote Green Risk Perrottet" (which sounds like an instruction to do both of those things) and fliers like the one above.

"This election will be incredibly close - to kick out the Liberals, Labor must win the seat of Balmain.

A vote for the Greens in Balmain risks returning the Liberals to power."

So, is there any merit to these claims?

What is the claimed risk?

One of the problems with voting scare campaigns like this is that they do not say exactly how the risk functions, and hence may mislead voters into thinking a risk exists that doesn't.  This particular flier might be read by various voters as meaning any of the following:

1. A vote for the Greens, even if Labor is preferenced, could assist the Coalition to win Balmain.

2. If the Greens retain Balmain they will not support Labor into government but will instead act in some way that assists the Liberals to remain in office.

3. If the Greens retain Balmain and fully support Labor then the fact that they have won Balmain will still result in Labor not forming government because of impacts on the decisions of other crossbenchers.

Meaning 1 is absolutely false as concerns Balmain (only the Greens and Labor are competitive) but there are a surprisingly large number of voters who don't understand preferential voting and don't realise that an excluded vote flows at full value, and some of these might be misled by such fliers.   (Incidentally meaning 1 is potentially true in Lismore, where there is a strategic voting issue surrounding Labor flows to Greens being weaker than Green flows to Labor in that seat).  

A further example backed off to "it could be impossible" for Labor to win without Balmain, before a third that spuriously raised the prospect of One Nation, who are unlikely to win any lower house seats.  I've seen a message from Labor candidate Philippa Scott that shows the intended meaning is in fact meaning 3, but meaning 2 is also worth looking at.  

Will this election be incredibly close?

It is common for politicians, parties and media to claim that elections will be incredibly close and that every vote matters whether this is actually true or not.   Some elections are very close, but a claim that a given election will be incredibly close in advance is overconfident.  Nobody can consistently forecast elections with that sort of confidence level, and elections that appear to be on a knife-edge often turn out to be not that close, or vice versa.  

The last 15 Australian state and federal (as distinct from territory) elections in a row have produced majority governments.  Twelve of those saw significant media speculation about a possible, probable or in cases supposedly inevitable hung parliament, but none of those hung parliaments happened.  At present it actually appears more likely than not (on aggregated poll-based modelling) that one will happen here, but that is no guarantee that it will.  For instance Labor's lead could increase, or Labor could get a favourable seat distribution and win a majority off its current aggregated polling (around 53% 2PP).  (A Coalition majority requires heroic assumptions but still can't be declared impossible on public evidence as at the time I send this either.)

More significantly, there are hung parliaments and hung parliaments.  The range of possible no-majority outcomes includes those where the winner is unclear on the night and negotations continue for weeks with government eventually being formed by a single seat, but it also includes plenty of outcomes where one side is very close to a majority, or at least in Labor's case needs the support of only one other component (the Greens) to get there.  As such the probability that Balmain is even capable of deciding the government is quite low; even if there is no majority, the effective margin will probably be more than one seat.  

Could Balmain make the difference?

Let's say that the numbers are such that the remaining 92 seats have not already ensured the result without considering Balmain.  There are a couple of possible scenarios where Balmain matters:

1. Balmain determines whether or not Labor is the most electorally successful party and some crossbencher ultimately bases their decision on which side to support or not on this.

By most electorally successful here I mean most seats, with two-party preferred as a possible tiebreaker.  So for this section to be activated, Labor would have to either lose the seat tally in a situation where a crossbencher would have chosen them had they tied it, or would have to tie the seat tally with the crossbencher choosing the Coalition.  For the latter case, this would almost certainly mean the crossbencher would be ignoring the 2PP vote.  

Labor figures here have made hay from both Alex Greenwich and Greg Piper stating that they would take into account most seats and "popular vote" (in Australia understood as 2PP) in making their decision.  However Greenwich and Piper have both flagged several other issues as key to their decision and so the chance of their decision being determined by electoral success is not that high. 

2. Balmain determines whether or not Labor can find a compatible group of crossbenchers to support it into government

The second scenario here is that Labor has enough crossbenchers willing to support it into government, except that some of them will not work with each other.  To give an artificial example, 44 Coalition, 42 Labor, 2 Green, 5 IND, where all of the independents prefer to work with Labor but three will only do so if they don't have to work with the Greens.  Here Labor can form government supported by the independents, but if the numbers are switched to 41 Labor, 3 Greens, they now can't.

This scenario might make sense if there was talk about establishing a coalition, but there isn't - all that is envisaged from Labor's side is confidence and supply from crossbenchers, not that those crossbenchers will work together on policy.  That alone makes this less likely to be an issue, but it's also unlikely that there will be that many crossbenchers who are Green-averse enough to make support from the Greens a dealbreaker.  

What would the Greens do in a non-majority situation?

The Greens recently launched a hung parliament strategy in a desperate attempt to frighten voters into running for the hills and electing a majority government.  Well that's how it always seems to me anyway - these displays of hubris from the party where it declares the election result in advance and starts trying to push Labor around before the votes are cast always strike me as unwise.  

Be that as it may, what the Greens said was that they would not support Perrottet's Coalition (some thought that left the door slightly open to supporting one led by Matt Kean) but that their support for Labor would be conditional on <insert long list here>.

Generally it would be expected that the Greens would find some way to support Labor into government and boast to their supporters that they had put Labor on notice on certain policies and perhaps extracted concessions on others (despite Labor's "no deals" pledge, tacit deals are very easy).  They've strongly hinted, but haven't outright guaranteed, that it would be something like that.

As the incumbent, Perrottet is entitled to remain in office until the parliament passes a no-confidence vote against him.  At this stage the Greens have not stated they would definitely vote in favour of such a vote, only that they would not vote against it, but this may be an unintended omission.  (At the federal election they had a clear position to throw the Morrison government out.)  The Greens have also not said they will necessarily vote Labor into office.  In theory if the Greens' position was to vote against confidence in Perrottet but abstain on confidece in Minns, a situation could arise where there were the potential numbers for a no-confidence vote in Perrottet, but not for Minns to survive a no-confidence vote himself, meaning that the former vote could be seen by other crossbenchers as pointless.  So it is possible to come up with a far-fetched scenario in which the Greens refusing to grant confidence to Minns could keep Perrottet in power, but it's an artificial idea that I don't think anyone really expects will happen.  Sooner or later the deal will be done, even if there is no deal.

Detractors of the Greens like to point to them supporting the Tasmanian Rundle Liberal government between 1996 and 1998, in order to imply they might do the same here (despite them ruling it out).  However, the situation of the Greens supporting the Liberals in Tasmania only arose because Labor refused to consider forming a minority government, meaning that the only choices were a Liberal minority or a second election.  In that parliament the Greens provided only the most minimal support on supply and confidence and frequently voted against the government they were supporting.   The Greens have supported Labor into government or to remain in government every time they had a chance to do so.

Overall

The flier's claim that Labor must win Balmain is unlikely to be correct and the flier's lack of precision makes it potentially misleading.  That said, this article has identified three scenarios in which Labor not winning Balmain could cost it government.  Each relies on a combination of two individually unlikely things happening together, so the overall chance that failure to win Balmain will cost Labor the election is remote.

Of course, should this occur, then every other seat Labor failed to win will have cost it the election too.  And some of those would almost certainly be a lot closer than Balmain.