There is a lot of talk at present about the melodramatically so-called "killing season", the time during the last week of November or early December in which struggling party leaders have good reason to be nervous. The theory is that this is the last opportunity to clear the decks before the parliament breaks for summer, and start the year on a new footing. But how much evidence is there that it is real?
The following list shows federal leadership bootings since 1970.
(I don't include cases where a leader resigned immediately after an election defeat, or apparently uncontentiously. A case could be made for including the Jan 2005 resignation of Mark Latham, who was certainly under leadership pressure, although officially resigning for health reasons.)
Before 1970, acrimonious leadership departures were rare on the federal stage. Menzies' essentially forced resignation as PM in August 1941 is one example, but there don't seem to be too many others.
I've also looked at this at state/territory level, but in that case there are data limitations in finding the true reasons for departure of a lot of leaders (without spending weeks buried in newspaper archives). In many cases we don't know whether their resignations were fully for the stated reasons, or whether they were really bloodless coups. Tasmania is a good example: did Paul Lennon jump or was he pushed by David Bartlett, and did Bartlett jump or was he pushed by Giddings? The public face of both of these cases was that the Premier simply resigned, but really we do not know what went on behind closed doors.
Looking just at state/territory leadership changes since 1970 I was able to find 25 that I classified as rollings against the incumbent's will (there were probably quite a few more) but of these only 5 conformed to the "killing season" narrative - Peter Collins (NSW LO, Dec 1998), Nathan Rees (NSW Premier, Dec 2009), Dean Brown (SA Premier, late Nov 1996), Robin Gray (Tas LO, Dec 1991), and of course Sir Johannes Bjelke-Petersen (Qld Premier, late Nov - early Dec 1987).
There are some cases (eg Rann in SA) in which the removal of an incumbent took some time to take effect from the decision to remove them. Counting the various rollings from the date of the decision being taken, the combined state/federal pattern looks like this:
Still, since three-quarters of bootings at state and federal level combined occur outside that window, it's not necessarily something to get too carried away about.
As far as Tony Abbott goes, there are some reasons not to get too carried away about prospects of his imminent removal. It might well happen - full-scale press gallery feeding frenzies can be savage things and it would not take much more to ignite one at the moment - but there are a few strategic obstacles. Firstly, the Coalition does not yet seem ready to answer the question of who to replace Abbott with, a question it has spent the last two years worshipping its 2PP and hiding from. Secondly with an election likely in the second half of 2013 (if Labor has the strategic wit to not repeat the blunders of Fraser and Carpenter, who called elections to try to capitalise on opposition leadership issues) it makes more sense for the Coalition to plan for a transition, say, three to six months out to maximise the "honeymoon effect" likely to result from the removal of an unpopular leader.
I predict that the surge will be larger than the usual three points should that occur, but the Coalition would want to go to the election with it still in force and surf to office on a national tide of relief that Abbott won't be PM, rather than with it starting to wash out of the system as questions about the replacement regime intensify. Thirdly it is not yet objectively clear that the Coalition's strategy in keeping such an unpopular leader has failed, only that it is so much riskier than it looked and would probably cost them an election that was close. We do not yet know if the next election will be.
The table above shows what may be a second spike in leadership removals, apart from the one in late November - early December. While there are not enough data to say for sure that it is meaningful, I suspect it is, because it makes intuitive sense. Struggling leaders who survive the end-of-year festivities unbooted but with questions hanging over their heads can usually last the summer, when not much is happening, but once the summer recess is over and battle resumes, the old questions soon enough resurface.
Those who last the "killing season" should beware the Ides of March?