The Pressure Valve: Does The Defeat Of Same-Party State Governments Help Federal Governments?

It's 1992.  The unpopular Cain/Kirner Victorian Labor government has been sent packing.  In comes Jeff Kennett and some voters are soon alarmed by his New Right agenda.  Cue massive protests.  The Keating federal Labor government has been struggling in the polls but it springs to life soon after Kennett's win (though that was far from the only cause).  In the 1993 election Labor gets a 4.34% swing in Victoria and gains four seats.  Across Bass Strait, where a short-lived Labor government had been removed in early 1992, there's an even bigger swing that yields another three.  The three Tasmanian losses are the first signs on counting night that something has gone terribly wrong with John Hewson's unloseable election, and these seven seats picked up by Labor in these two Liberal states combined are the backbone of Keating's against-the-odds win.  

Victoria 1992 is the paradigm case for a theory that one might call the "pressure valve" theory of state elections, that there is a drag effect of state elections upon federal elections and that federal governments benefit if the voters let off steam by throwing out an unpopular state government of the same party instead of taking their anger with it out on the feds.  Better still if the new state government has started to frighten the horses.  I have talked a lot about "federal drag" on here, which refers to the fact that state governments do much worse at elections, all else being equal, when the same party is in power federally.  Age and federal drag are the two biggest killers of state governments and it is for this reason that the Miles Government was always likely to lose by about as much as it did.  But does it work the other way?

Many years ago - and I am not sure it is still online anywhere - Matt Cowgill posted some analysis of "state drag" that found, as best I remember, that if "state drag" on federal elections existed at all it was much much weaker than federal drag, and that even establishing clearly whether there was something there at all was rather complicated.  Remembering this I have generally not seen much in the theory of "state drag" when people have asked me about it.  But as this very specific version of the "state drag" claim is being talked about a lot in the wake of the dumping of the Miles Government, I thought I would look at it and see if there is anything in it.  

The first simple test I looked at is this: if a same-party state or territory government loses office, is the federal government's 2PP swing at the next election better or worse than the national 2PP swing, and if so by how much?  

For the last 50 years I found 28 cases of a same-party state or territory government being voted out of office: three in NSW, five in Victoria, three in Queensland (I have included 1995-6 as the effect of the 1995 election plus the 1996 Mundingburra by-election was that the Goss government was removed before the 1996 federal election), six in Western Australia, four in South Australia, just two in Tasmania, three in the NT and two in the ACT.  

Between 1974 and 1987 there were not that many cases of same-party state governments being kicked out though Coalition governments in SA, Victoria and WA were all thrown out in the leadup to the Fraser federal government being likewise evicted in 1983.  From the 1987-1990 federal term on at least one same-party state government has been defeated in the leadup to every federal election except 2007 (the reason in that case being there were none left to defeat after three terms of Howard federal drag.)  The most such cases happened in the 1998-2001 term (Coalition governments in Victoria, WA, NT and ACT) and the 2010-2013 term (Labor governments in NSW, Victoria, Queensland and NT).  The 2010-2013 case is exceptional in that 78.4% of the population lived in areas that had kicked out their state government during the federal term.

What I find on that basis is ... nothing to see!  Of the 28 cases, in 14 the federal government did better than the national swing after the defeat of a state government.  In the other 14 cases it did worse.  Overall in the mean case the federal government did better, but only by 0.14 points which is not even remotely statistically significant as the standard error on this estimate of the mean is 0.36 points.  

Having mentioned two 1993 cases in point for the "pressure valve" theory it is worth looking at the major examples against it.  In 1983 the Liberal government in Western Australia had been evicted during the federal election campaign but this did not prevent the swing to Labor in WA being 4.6 points higher than the national swing (I am using David Barry's estimates, others are slightly different).  A possible special factor there is that Bob Hawke had strong connections to Western Australia; although born in SA he had grown up in WA and his uncle had been a Labor Premier there.  

The end of the Goss government across 1995 and 1996 did not stop the swing against Labor in Queensland in John Howard's 1996 landslide win from being 3.6 points higher than the national swing.  Again this government was defeated only during the 1996 campaign and with the novelty of a mid-term change of office depending on the vote of a crossbencher, it may not have really sunk in that there was a lasting change at state level.

The third largest case on the negative side also comes from Western Australia and also involves a very recent change of state government at the time of the federal election.  It is in fact 1993, when the decisive but 'not that bad' loss for Labor at state level encouraged Keating to call the federal election right away.  Labor in Western Australia underperformed the federal swing by 2.65%.

There's an argument here that when the state government has only just lost office at the time of the federal election, there might not be time for any effect to kick in yet.  The new state government might be undergoing a honeymoon effect that might lift the federal Opposition's brand enough to cancel out any relief factor at the booting of its predecessor, for instance.  

If I rerun the test to exclude all cases where the gap between the state and federal elections was less than three months, the theory does considerably better.  Now there are 12 cases of outperformance to eight of underperformance, and the mean swing difference is 0.74 points (the median is 0.45, with the Tasmania 1993 outlier having a big impact on the mean).  So there could be something in the "pressure valve" theory but (i) if so it only kicks in once the new state government has had time to do stuff some voters might not like (ii) if there is an effect, it's not a big one.  

There are only seven cases where the party that would benefit from the effect has overperformed by over 1%.  Three of these are in Tasmania and the NT, and of the others Queensland 2013 could have had more to do with the re-instatement of Kevin Rudd (Qld) as Prime Minister, and NSW 1990 could have been influenced by the dumping of John Howard (NSW) as Opposition Leader.  The remaining cases are Vic 1993 and 2016.

Why isn't state drag a more prominent thing?  One view I have on this is that state politics is seen as subordinate to federal and so voters are far more likely to use a state election to sound off at a federal government.  They want a state government that will stand up to the feds but not enough to kick out the federal government just to achieve it.  Another is that from time to time a particular issue overlies state and federal politics in a given state and makes the federally governing party both more likely to suffer a swing against it at both levels and hence more likely to lose.

As I often hear it, the "pressure valve" theory also argues that throwing out an unpopular state government doesn't mean that the federal government will do well in the state, just that it will do better than it might have done if that state government had not been thrown out, ie that unpopular state governments drag the same party's government at federal level.  NSW and Queensland had high swings against Labor in 2010 and also had very unpopular state Labor governments at the time (in Queensland's case the dumping of Kevin Rudd was a factor as well).  Some other possible examples are Victoria in 1990 and Tasmania in 2013.  

I haven't tried to evaluate this one yet as the state polling record is very patchy and also as state government popularity is often influenced by federal factors anyway.  Often when a same-party state government does get the boot, its polling declines in its last year and may not have necessarily been as bad more than a year out.  So I think it would be challenging to get enough polling data - especially from the last 10 years or so - to judge this one from.  

At the least, my analysis suggests that Labor being thrown out in Queensland is unlikely to be a path to a large overperformance for Labor in the state.  

Note: Because the national swing includes the swing in the state(s) that evicted their government(s), the difference between the swing in those states and the overall swing in all other states, as opposed to the national swing, is going to be slightly greater on average than the numbers I give.  It's not obvious what is the best way statistically to test differences - for instance because 2013 saw four same-party state governments booted, the large swing to the Coalition in Tasmania 2013 gets over-represented.  I may at some stage look at overall swing in states where a same-party government was booted vs overall swing in states where this did not occur, by election.  In 2013 the difference in this case was 0.60 points.