I was interested to see if we're actually seeing more party leadership changes in politics these days, or if it just feels that way, so I knocked this app up quickly to look at how often changes happen historically. I've only looked at changes of the head of government here, not changes in opposition parties.
From the options below you can choose a date range and a set of states/territories/the commonwealth to look at, and then the app will show you graphically whenever someone started a term as head of government. There's also a list of all those people with their term lengths underneath. People who served as leader at two distinct times (eg Kevin Rudd) are counted as having two separate terms of office.
There is the option to include currently serving leaders, but note that this will distort average term lengths.
To crudely test if term lengths are actually shorter than historically normal I'm just going to look at the average over 20 year intervals going back to federation (excluding currently serving leaders). I've also included the last ten years as an overlapping interval. Remember that each leader is counted in the interval that they started their term.
|Time Interval||Average Term Length|
|1 Jan 2005 - 31 December 2014||940 days|
|1 Jan 1995 - 31 December 2014||1523 days|
|1 Jan 1975 - 31 December 1994||1359 days|
|1 Jan 1955 - 31 December 1974||2113 days|
|1 Jan 1935 - 31 December 1954||1619 days|
|1 Jan 1915 - 31 December 1934||1063 days|
|1 Jan 1901 - 31 December 1914||730 days|
So yes, from this crude measure, it's credible that we're churning through leaders much faster than historically usual.
It's interesting that around the time of federation it was entirely normal for Prime Ministers to serve very short terms, and often get two or even three separate goes at the job. It also seems our leaders hit their longest terms in post World War Two years.
It's hard to see any speeding up of the rate of leader churn in the plot below, but it's interesting to play with nonetheless.
Let me know if you see anything interesting in this dataset in the comments or on Twitter @CaseyBriggs
Update 4 Feb 2015, 11:48am: I just discovered an error in the app that was excluding Western Australian leaders (I accidentally inserted a rogue space in just before I deployed it!). That's fixed now and I've updated the table of stats above. Most of them changed by less than 10 days, one changed by about thirty, and one changed by about 100.