Ross Church’s victory has been portrayed in some quarters as a resounding rejection of two term incumbent Jenny Rowan. A careful look at the first preferences and iterations shows that this has been exaggerated.
I also argue that the way the results are presented – for example by the Dominion Post as well as one KIN article – is misleading.
Most reports give the figures as follows: Church 7,933, Guru 6,779, and Rowan 4,543 (and less for the other four candidates). But this is not comparing like with like.
Under the Single Transferable Vote (STV) system, the lowest polling candidates are eliminated one by one until one candidate has over 50% of the votes still counting. As candidates are eliminated, the second preferences of those voting for them are used – and even third preferences and beyond.
‘Jenny Rowan lagging much less’
After the lowest polling four were eliminated in this election (iteration five), the votes were Church 6676, Guru 5,666, Rowan 4543. This is the fairest comparison of votes to make – and it shows Jenny Rowan lagging much less than the result above.
At this point Ross Church does not have a majority, so Jenny is eliminated and her second preferences re-allocated to yield the final result for Ross and Guru in the first paragraph – iteration six.
Slightly more of Jenny’s first preference votes went to Ross than Guru – and nearly half went nowhere. Those people either voted only for Jenny or voted second for candidates already eliminated.
For all five candidates eliminated, the published results show only their first preference numbers, whereas for the two involved in the final choice, they show much higher figures than their first preferences, due to these re-allocations – hardly a fair comparison.
Rowan gets more than Guru on first count
It is also interesting to look at these first preferences only – these were Church 4582, Guru 4016, Rowan 4069, with Welsh next at 2,858 and a total for the seven candidates of 19,046 (so over 9,500 first preferences would have been needed to be elected with a clear majority straight off).
The votes were well distributed, with leader Ross Church gaining 24.4% of first preferences.
It is also notable that Jenny Rowan had slightly more first preference votes than Guru and that all three were not that far apart.
Ross was ahead at each iteration and gained more second preference votes than the others so the gap widened. He would of course have also won under First Past the Post.
STV reflects voters’ feelings about several candidates
However, one important aspect of STV, which I see as a virtue, is shown by this election – it reflects voters’ feeling about a number of candidates, not just their top one. It gives more confidence in the election result as the gaps widened with reallocation so less voters are unhappy with the result than with any alternative.
It is clear that on the whole voters either strongly wanted Jenny Rowan to have a third term or strongly wanted a change – perhaps not strange for an incumbent.
I say this because she gained under 500 reallocated votes between iterations 1 and 5, whereas Ross Church gained just over 2,000 and Guru over 1,500.
It is also interesting to note that in the last iteration, Ross’s 7,933 votes plus Guru’s 6,779 total only 14,712, with 7,357 needed for a majority, over 2,000 less than needed on first preferences – I’ve explained how this happens with the reallocation numbers for Jenny’s votes above.
Be thankful I have confined myself to the mayoral election!
It gets more complicated!
STV is much more complicated when electing more than one person – as with the Councillors and Community Boards. Reallocation of surplus votes for those elected come in.
The only one I looked at in detail was the Paekakariki Community Board (my home) where four were to be elected from five candidates.
In that election, three had sufficient votes to be elected on first preference (with Jack McDonald top), yet it took six iterations to separate the two bottom candidates.
So well done, Philip Edwards, and bad luck, Graeme Tuckett – at the sixth iteration, believe it or not, 122.7994461 votes were needed (!) to be elected. Philip had 122.95, Graeme 120.42.
Elections can be strange beasts!