To paraphrase Mark Twain, "it's not the percentage of the swing that matters, but the number of votes in the swing." I've taken a stab at predicting the final breakdown of party votes in each electorate, and should these results prove accurate, it supports the general narrative of this election that while Labour did well around the country and in the inner cities, it failed to deliver the knockout blow it needed in West and South Auckland.
There's a few interesting things that emerge when you look at how many extra votes per electorate have been shifted from the centre-right to the centre-left. The first is that special votes hold true to 2014, there will be a swing to the centre-right in just two seats - New Lynn and Manukau East. If you check out the assumptions at the bottom of this, my gut feel is that New Lynn will wash up as a small net swing to Labour in the end, but Manukau East could still hold out.
It's also interesting to see that the bulk of the lowest contributors to the centre-left swing are again West and South Auckland, reinforcing how poorly Labour did there. Certainly, Labour dominates the share of the party vote in those electorates, but they generally have pretty poor turnout rates, and without Labour simply seems to have been unable to mobilise any additional support in those electorates.
Where Labour really did capitalise was across inner city Wellington, Christchurch, Dunedin, and Auckland, as well as some big provincial centres too like New Plymouth, Nelson, and Whangarei. With the average swing being 3,760 votes, everything north of Taranaki-King Country is a win for the centre-left, and there's a good representation of provincial cities and rural electorates there.
What was also crucial about these seats is that most have high turnouts, meaning that the impact of any swing is magnified. For example, Selwyn may have had a 12.44 percentage point swing to the centre-left, but because Selwyn is likely to have the third highest turnout, they've contributed more to the overall swing to the left that Mt Albert, Christchurch Central, or Auckland Central (which had the largest non-Māori electorate swing of all) who all had larger swings, but smaller voter turnouts.
While all the Māori electorates delivered big swings to the centre-left, the lower voter numbers in these seats means that put within context of the overall shift, they're relatively evenly spread with four delivering above average swings.
The big opportunity for the centre-left bloc over the next three years is to crack the turnout issue in West and South Auckland and turn it to their advantage. With big majorities, and generally much lower than average turnout in those seats, if they can finally figure out how to get those voters out in force, they'll be able to dominate the electorate for years to come.
There's a few assumptions underpinning this prediction. The first is that each electorate receives special votes at roughly the same share of the overall total of special votes as they did in 2014, and that the allocation of those special votes roughly follows the same pattern as 2014 - that is that National and NZ First drop 17%, Labour gains about 14%, and the Greens gain 53% (my hunch is that Labour and the Greens will be much closer on these figures than the 2014 split, and National and NZ First will do worse overall, but this is the best guess we have currently).
Also I've defined the swing as the movement in party vote share per electorate from the Government bloc (generally National, but in the Māori electorates it's National plus the Māori Party) to the joint Labour and Green Party bloc. I've kept New Zealand First out of any swing calculations as they're capable of going either way to help form a government, so it's not really possible to claim their share of the party vote as part of any swing.