Mitt's 2012 Mirage
Want To Understand Politics? Give 2012 A Miss
One of the bigger surprises two straight Australian elections was the seat of Longman. In 2016, the LNP lost it despite a 7% margin (in Australian terms, a 57-43 seat after preferences is on a 7% margin - no, I don't write the rules), a shocking result. After a 2018 byelection saw a sizable swing to Labor, everyone and their brother thought Labor was home and dry in the seat, such that you could get the LNP at 6/1 to win the seat the day before the election. And then they won it with a sizable swing back, returning the seat to a comfortable margin with the LNP at just over 53%, after preferences.
Down a couple of states, you hit the Melbourne seat of Kooyong, held by the now-Treasurer Josh Frydenberg. In 2013, he won it 61/39, a lead which expanded to 63/37 in 2016 despite a national swing against the Coalition. Then, in 2019, despite a national swing to the Coalition, his margin took a sharp fall - down to 56/44.
What's interesting about both Kooyong and Longman is that the 2019 results are not particularly surprising given where they were in 2013 - the Coalition bled a couple of points nationally, so it's unsurprising both seats were more red after two terms in office than at their inauguration after 6 hard years of Rudd/Gillard/Rudd. However, both results look more surprising in the context of 2016 - Frydenberg's seat looking more safe than it really was, and Longman obviously flipping for a term before coming back. The thing is, with hindsight, it all makes sense.
Longman is a regional seat up in Queensland, where energy jobs are plentiful and the concerns about climate policy aren't broadly shared. Kooyong, by contrast, is a wealthy seat in Melbourne where posh social liberals live, and where tax cuts and responsible action on climate play well. They're two entirely different types of conservative seats, and they swung very differently. Why? Because in 2013 and 2019, the Coalition was helmed by social conservatives whose belief in climate change could be considered, uh, derisory. Both Tony Abbott and Scott Morrison don't actually care about climate policy, they care about winning power - and if that means chucking climate policy under the bus, so be it. Abbott might actually be a denialist, as opposed to merely an opportunist, but in terms of policy positioning, the Australian right had good fits for Longman in their leader in both 2013 and 2019. In 2016? They had Malcolm Turnbull, a posh social liberal who lost the leadership in 2009 for agreeing to support the Rudd Government on a price on carbon and believed fervently in gay marriage as a good and just thing. In hindsight, it seems fairly obvious that Turnbull would struggle in Longman and seats like it, but more than holding his own in seats like Kooyong. And, once Turnbull was out of the race, it shouldn't have been surprising that a more culturally conservative approach would do better in regional Queensland, but fail in the inner cities. This should be less shocking than it was, both to the general consensus, but also to me.
…
Head to the US, and you see a similar pattern in the Nebraska 2nd. Obama won the current NE-02 by just over a point, which blew out to Romney +7 in 2012. Since then, it's gone from there to Trump +2 and then Biden +6.5, despite almost no national swing from 2012 to 2020. Head to the Iowa 2nd and you see the opposite - Obama won it by 15.5% in 08, and held it by 13% four years later - a better result for Democrats relative to the nation than 2008. Trump would win the district by 4% in 2016, a result that would be matched against Joe Biden four years later.
In both cases, you see a mirage - 2012. 2012 made the Nebraska 2nd look much more Republican than it was, and that 2012 Iowa 2nd result would have been rather frightening for Republicans looking to win the state back. Now, the 08-16-20 trend line makes sense, especially against national environment. Hillary outran Obama in the Nebraska 2nd adjusted for the national environment, and then Biden stormed home there, and there being some slippage in the Iowa 2nd - both because of national environment and decaying rural support for Democrats.
If you want more examples, there are plenty - the Kansas 3rd went McCain +1, Romney +9.5, Clinton +1, Biden +10.6, and Cheri Bustos' IL-17 is a spitting image of the Iowa 2nd. Romney was a tremendously good candidate for the suburban areas and a horrendously bad one for rural and regional America, and therefore our perceptions of how the last 8 years have gone were warped. We all collectively thought regional areas like Scranton were more Democratic than they really were, and that Joe Biden could do a lot better there than Hillary did. The problem was, we had the halcyon days of 2012 in our minds, which wasn't replicable. That was about Obama, sure, but it was really about Romney.
Look at this properly, and you see that the results today are less shocking if you start in 2008, because 2012 was the aberration, the exception. Thinking of it as anything other than that is a mistake. In the same way that it should have been obvious that Turnbull would do much better in the cities and struggle in rural Australia, it should have been obvious that running Not Mitt Romney in 2016 would have snapped some of the rust belt back to a less Democratic partisan baseline. Unless you think the GOP are nominating a Romney-esque figure again, which of course they won't, then 2012 merely adds noise to your search for signal.
The other thing that needs to be said is that places can snap fast - either left or right. We think of realignments as smooth processes, but they're not. They're fast. Longman went from 57/43 to 49/51 in one term, and then went back to 53/47. Frydenberg went from getting 63% of the vote to 56% in 3 years in Kooyong. Georgia 6th has moved 30% from the 2008 baseline to 2020. These seats can move, and move rather fast, when the voters decide to move. This is not a prediction that any places will snap, but just a reminder that when things fall apart, they really fall apart.