Home > Uncategorized > More Matt McCarten Meditation

More Matt McCarten Meditation

October 30, 2010 Leave a comment Go to comments

Here’s a comment from Markus that further attempts to pour cold water on my low-balling of Matt McCarten:

    But while I think many are grossly over-estimating Matt’s chances, I also think 200-400 votes is way too low, Phil. I can see him coming a close fourth behind the Greens’ Logie, possibly even just sneaking into third place. Matt is pretty well-known and should receive enough support from a (relatively small) minority of Labour voters and a (somewhat larger) minority of Green, NZ First and minor-party voters to take maybe 1300-1600 votes. But that would still leave him over 9000 votes behind the first and second-placed candidates.

This is food for thought but I am sticking to my guns. The biggest flaw in Markus’ analysis is his over-estimation of the number of votes in play.

The two by-elections in general electorates under MMP (Taranaki-King Country 1998, Mt Albert 2009) suggest turnout will drop to around 20K from approx 35K at the 2008 general election, a 43 percent drop. Roughly, this means Labour’s base vote (the 2008 figure minus 43%) is 8,500, National is at 7,400 the Greens stand at 1,600.

This just goes to show how outlandishly exaggerated the 6,000-vote majority figure is, as touted by National Party spin-doctors on Kiwiblog and elsewhere.

The real majority is closer to 1,000 — and that is making the heroic assumption that the fall-off in voter turnout will be even across the electorate. If the recent local body elections are any indicator, this will not be the case at all; the Labour-friendly east and west is likely to stay at home in greater numbers than the Tory north. Anyone who follows elections in Mana/Porirua will simply factor in such a turnout discrepancy as a matter of course.

For Matt McCarten, this means there are simply fewer votes in play than Markus imagines. Indeed, by subtracting for by-election turn-out, Markus’ projected McCarten vote drops to 750 before we even begin to argue.

I maintain that MM is not a vote magnet in Mana, any more than Russell Marshall was in the recent Mayoral election (he came fifth). I further maintain that McCarten does not benefit from a favourable electoral climate; thus, his candidacy lacks a compelling narrative. Beyond the small minority of voters who know enough about McCarten to support his quixotic efforts out of ideological affinity or affectionate nostalgia, there just isn’t any good reason why anyone would vote for him.

Finally, his “chicken-in-every-pot” policy manifesto is irrelevant, and will go nowhere, since he promises to occupy the most distant and powerless corner imaginable in the Parliament.

Categories: Uncategorized
  1. markus
    November 4, 2010 at 00:38

    Interesting argument, Phil. (And I always like to see my name up in lights).

    A considered response should be winging its way towards this site hopefully by about Sunday/Monday (depending on how much spare time I have). Very interesting stuff here on the mid-term elections, as well.

    – Markus.

  2. markus
    November 5, 2010 at 22:07

    My response, Phil:

    (1) To the idea that turnout figures in the (1998) Taranaki-KC and (2009) Mt Albert by-elections suggest turnout in Mana will, likewise, drop by 43 % to about 20K (and that therefore far fewer votes will be in play for Matt McCarten than I imagine).

    (A) Even if we were to assume that your 43% figure is correct, your extrapolation from this of probable party base-votes in Mana is surely wrong ? According to my calculations, Labour’s would in fact be 8700 (not 8500), National’s 7200 (not 7400), and the Greens’ 1700 (not 1600). Hence, the two parties (Lab/Greens) most likely to contribute to McCarten’s vote would have a somewhat larger base-vote than your figures suggest.

    (B) Regarding Mt Albert, your turnout figures seem mistaken, effectively over-emphasising the degree of decline. 34,963 Party-Votes were cast in Mt Albert in 2008 and 20,885 votes were cast in the 2009 by-election. That represents a 40% (rather than 43%) drop. In terms of Mana, such a drop would mean base-votes of 9100 Labour / 7600 National / 1700 Green. So while you argue that the Mt Albert and Taranaki-KC turnouts point towards a combined Labour/Green base-vote in Mana of 10,100, my calculations suggest it’s in fact a rather larger 10,800.

    • PQ
      November 5, 2010 at 22:16

      I am sure you are right. These are rounding errors, though. I stand by my larger point.

  3. markus
    November 5, 2010 at 22:28

    But much more importantly:

    (C) Far from closely coinciding (and thus being strongly indicative of what will happen in Mana), the turnout figures for these two by-elections are in fact quite disparate.

    Sure, there’s a close fit in terms of raw numbers – 20K in Taranaki-KC / 21K in Mt Albert – but, crucially, the turnout PERCENTAGES diverge greatly. And surely in the context of two by-elections conducted more than a decade apart (with the significant increase in eligible-voter population that this entails – 33,000 on T-KC master roll in 1998 / 44,000 on Mt Albert’s master roll in 2009), it’s the turnout PROPORTION (rather than raw number) that’s the key figure here ?

    While the 2009 Mt Albert turnout was down 40% on 2008 (20,885 compared to 34,963), the 1998 Taranaki-KC turnout was down only 26% on 1996 (20,225 compared to 27,504).

    You can’t therefore argue that 1998 T-KC points toward even a 60%, let alone 57%, turnout in Mana. It in fact suggests a 74% turnout. (***obviously in this particular context I’m using “turnout” to mean by-election numbers as a percentage of numbers turning out in the previous general election – 20,225 is 74% of 27,504 – rather than percentage of all those on the electoral roll).

    So, I would argue that we in fact have two completely divergent by-election turnouts here: 60% Mt Albert (47% of all eligibles) / 74% T-KC (62% of all eligibles).

    • PQ
      November 5, 2010 at 22:31

      This is forensic. But Mana will be closer to Mt Albert than T-KC, disagree?

    • PQ
      November 5, 2010 at 22:41

      I made a stupid error I think, by looking at the 2008 General Election figures for Taranaki-King Country when I should have looked at 1996 numbers.

  4. markus
    November 5, 2010 at 22:40

    So, what does this divergence mean ?

    That by-election turnout is completely arbitrary and unpredictable ?
    That Mana’s turnout will be somewhere between the two ?
    That by-election turnout is much lower in urban than rural seats ?
    That turnout is invariably lower nowadays than in the recent past ?
    Or that one of these two by-election turnouts should be considered an anomoly ?

    It’s hard to say for sure (and we’re obviously going to find out soon enough), but I think there is at least some evidence pointing towards the last of these options – with the 2009 Mt Albert turnout being the anomoly.

    (Note: As you can see, Phil, I’m sending this response in a series of comment-boxes. My clapped-out computer with low-memory just can’t manage to deal with sending longer messages all in one go. So, there should be two or three more of these comment-boxes below over the next hour or so).

  5. markus
    November 5, 2010 at 23:00

    Consider this:

    (a) First – and least important – there appears to have been widespread surprise in the blogosphere at the low turnout in Mt Albert (47% eligibles / 60% of 2008 turnout). National’s pollster David Farrar, for instance, described it as “very low”.

    (b) More importantly, it’s clear the National vote absolutely COLLAPSED. Just look at the numbers:

    Mt Albert 2008 Party-Vote /////Mt Albert 2009 By-Election/////Difference

    Labour 14,894———————13,260————————–minus 1634

    National 12,468———————3,542————————–minus 8926 !!!

    Green 3846————————-2,567————————–minus 1279

    Act 1227—————————–968—————————minus 259

    The Labour candidate’s 2009 vote represented 89 % of Labour’s 2008 Party-Vote / the relevant figure for Act was 79% / the Greens 67% / while National’s Mellisa Lee managed a mere 28% of the Nats’ 2008 Party-Vote. An astonishing collapse !

    You’ve argued, Phil, that while National traditionally manages to mobilise its support-base, Labour has always had much more difficulty – anything else, you claim, represents a “heroic assumption”. In which case, you too must surely view this utter collapse of the Nat vote as abnormal ?

    (Note: Hope you don’t mind if I continue sending the rest, before responding to your replies ?).

  6. markus
    November 5, 2010 at 23:27

    At the outset of the campaign, National appeared to have relatively high hopes. On a Red Alert post, Trevor Mallard suggests Nat pollster Farrar was calling the possibility of a narrow win for National just before Mt Albert candidate selections. There were apparently suggestions that polling was very close.

    The only Farrar prediction I’ve managed to find (after a very brief search) on Kiwiblog is his May 5 2009 analysis based on the 2008 Party-Vote, adjusted for changed nationwide party poll ratings and (his) estimation of likely tactical-voting patterns (not entirely unlike our analysis here). So, I’m not sure if any private National Party polling actually took place in Mt Albert, but certainly based on this number-crunching, Farrar decided a narrow National win (by a few hundred votes) was at least possible.

    Then, of course, Mellisa Lee (apparently unpopular with National activists right from the very start) indulged during the campaign in that extraordinary faux-pas frenzy which (perhaps together with SuperCity discontent / a bad TVNZ poll for the Nats / a belief among National supporters that a win was impossible) led to this massive collapse.

    The point being that arguably 7000-8000 National voters appear to have stayed home (allowing for relatively small Nat-to-Lab and Nat-to-Green swings). Under different circumstances and with a better candidate, most (5000-6000) of these National supporters would surely have voted ? (the other 2000-3000 being the number of Nats you’d expect to stay home regardless of the circs).

    I guess it’s possible to argue that although David Shearer’s 2009 vote was only 1600 down on Labour’s 2008 Party-Vote, most of this may in fact have come from 2008 National supporters. (the idea that many erstwhile Labour people did indeed stay home – as one might expect – , that Labour’s vote only held up – and National’s declined so disasterously – because a huge number of Nats (it would need to be a large majority) swung to Shearer, and therefore that Mt Albert’s overall turnout was not, in fact, abnormal).

    Personally, I think that’s very unlikely. I just can’t see a majority of Nats voting Labour under any circumstances. Even PM Helen Clark only managed to take about 2000 Candidate-Votes from National’s Party-Voters in 2008. ‘When two tribes go to war’ and all that…

  7. markus
    November 5, 2010 at 23:53

    The evidence, then, suggests an abnormal / exceptional collapse of the National vote and that under normal circumstances we might expect those 5000-6000 National supporters to have turned out.

    Add these 5000 votes to the 2009 Mt Albert turnout and you get: 25,885 (74% of the 2008 turnout – precisely the same as T-KC’s in 1998).

    So, I’m suggesting we should in fact expect the 2010 Mana turnout to be about 26,000, roughly three-quarters of the Mana 2008 turnout (about 60% or so of eligibles, down from 82% in 2008). Your estimation, I’m arguing, is too low.

    This takes us on to your argument that by subtracting for by-election turnout, my “…projected McCarten vote drops to 750 before we even begin to argue.”

    Clearly, then, you assume that my 1300-1600 estimation is premised on the idea that 2010 by-election turnout in Mana will be about the same as in 2008 (750 is roughly 57% of my 1300 minimum). But that is not, in fact, the case. I’ve assumed 2010 turnout will be about 75% of 2008 turnout (for the reasons outlined above).

    Let’s look at the figures I gave you:
    McCarten (Ind) 1300-1600 votes
    Logie (Green) about the same as McCarten or a little higher
    Fa’afoi (Labour) and Parata (National) both more than 9000 votes ahead of McCarten.

    Those figures in fact dovetail perfectly with my 75% turnout prediction.

    The following isn’t my final prediction (regarding the Fa’afoi and Parata vote), but it gives you a ball-park idea of where I’m coming from:
    Fa’afoi (Lab) 11,900
    Parata (Nat) 10,700
    Logie (Green) 1600
    McCarten (Ind) 1400
    All others combined 400
    Total vote 26,000 (75% of the 2008 Mana figure). McCarten is over 9000 votes behind Parata (9300 behind) and Fa’afoi (10,500 behind).

    Even if we accepted your lower turnout figure (60% of 2008), my projected McCarten vote would therefore only drop to about 1100-1400 (because I have already factored a 25% turnout drop into my calculations).
    Now, I haven’t quite finished my response – I still want to look at turnout differentials and respond to your replies – but we’re about to go out for the evening, so it’ll have to be tomorrow. Thanks for indulging me (and my increasingly eccentric computer), Phil.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: