Wednesday, April 19, 2006


Below is the first draft of my submission on Barbara Stewart's Electoral (Reduction in Number of Members of Parliament) Amendment Bill. It is based primarily on arguments here and here. While submissions haven't actually opened yet (the Justice and Electoral Committee hasn't set a date), I thought I'd get in early and refine my thoughts a bit.

  1. I oppose the Electoral (Reduction in Number of Members of Parliament) Amendment Bill for the following reasons:
  2. The bill would reduce the overall size of Parliament by cutting the number of list seats without any corresponding reduction in the number of electorate seats. This would result in more overhangs and fatally undermine the proportionality which is the goal of MMP.
  3. Under the current system, overhangs are the exception rather than the rule, and the usual beneficiaries are small parties. The current situation (where the Maori Party has one seat more than it is entitled to) is the first time this has happened since MMP was introduced in 1996. But if the bill passed, electorate seats would make up a far larger proportion of the House – approximately 70% as opposed to 58%. The FPP system used for electorate seats means that they are not distributed proportionately; rather, differences in vote share between parties are magnified, and the largest party gains a disproportionate share. Currently, this does not matter. A party would have to sweep almost all of the electorates in order for overall proportionality to be affected – something which has never happened in the history of party politics in New Zealand (the largest landslides – in 1905 and 1990 – saw the winning party gain just over 70% of the seats). But in a smaller House with the same number of electorates, it is far more likely that this magnifying effect would affect proportionality and result in an overhang for the largest party.
  4. For example, if the 2002 election had been run under the system proposed in the bill, Labour would have gained 45 seats on 41.26% of the vote – 2 more than they would have been entitled to (the difference being due to the “wasted vote” – votes for parties which did not make the threshold). National would have gained 22 seats on 20.93% of the vote, and gained only one list seat. If the 1999 election had been run under this system, the 41 electorate seats won by Labour would have been exactly what it was entitled to, and it would have gained no list seats. Only in elections where the two major parties were very closely matched (as in 1996 and 2005) would there be a low chance of an overhang.
  5. More generally, if the largest party gets 40% of the vote, it would have to win only 57% of the electorate seats before it affects proportionality. This has happened in seven of the twelve elections since 1972. While the effect will not be as great as it was under FPP (the losing party gaining a top-up from the list), this suggests that overhangs benefiting the larger party would once again become an entrenched feature of our electoral system.
  6. This problem will only get worse, as the number of electorate seats will continue to grow at the expense of the list. Reducing the number of list seats so greatly will result in progressively more and more overhangs until proportionality completely breaks down (something expected to occur when the number of electorates reaches about 75% of the total).
  7. In addition to providing regular overhangs for the largest party, the bill would also magnify their effect, giving their recipients proportionately more power. One extra seat counts for more in a 100-seat Parliament than in a 120-seat one. Thus they will be more of a distortion, and there would be more of an incentive for parties to deliberately seek to “game the system” to benefit from an overhang. This would reduce the legitimacy of the electoral system overall.
  8. One solution would be to amend the Electoral Act to reduce the number of electorates proportionately to the overall reduction (i.e. to 58 seats). This would require reducing the number of South Island seats to 13. Quite apart from the legal difficulties (the appropriate clause of the Act, s35, is entrenched and change would therefore require a supermajority or a referendum), this would likely result in electorates which are simply too large to be practical.
  9. In short, the goal of a reduction in the number of MPs cannot be achieved without fatally compromising MMP. While some would point out that a smaller House was approved by an indicative referendum in 1999, it is worth remembering that the move to a proportional system was also approved by a far more rigorous and binding referendum process in 1992 and 1993. It would be inappropriate to overturn the results of that referendum without a similarly rigorous process which ensured that the electorate was fully informed of the consequences.
  10. As a final note, the increase in the size of the House under MMP has not just ensured proportionality; it has substantially reduced the power of the executive and given us better oversight by Parliament. Many commentators have noted the increased effectiveness of Select Committees under MMP, and this is no accident. More members has meant more people available to do the work of scrutinising legislation, rather than simply rubberstamping it. It has also meant more eyeballs watching the government, ensuring a higher level of accountability in general. Reducing Parliament to 100 members would return us to the pre-MMP days of an unaccountable executive lording it over a rubberstamp legislature. I do not believe this is desirable, or in fact desired by the majority of New Zealanders.
  11. I do not wish to make an oral submission to the Select Committee.


Excellent. You make your points well.

Posted by Lewis Holden : 4/19/2006 09:56:00 AM

Great. I have the following comments, which I might put into a submission of my own:
1. The financial saving from a reduction to 100 members is likely to be in the order of $4-8mln (based on the 2005 Parliamentary Service vote). This represents just one gold coin for each New Zealander.

2. The referendum question did not inform the voter of either the negative effect on the representative nature of parliament or of the miniscule cost savings resulting from a reduction in the number of MPs.

3. Reducing the number of MPs would reduce the pool of candidates able to become Ministers. This would have the undesirable effect of limiting the size and competence of the ministry and hence reducing the level of democratic accountability applied to the public service.

4. Finally, I would suggest that the electoral system should not be modified lightly by a parliament. Although there is no legal obstacle to reducing the number of list (as opposed to electorate) seats, I would suggest that *any* substantial change to our arrangements should be debated in a proper constitutional convention and approved by a statutory referendum.

Posted by Rich : 4/19/2006 12:01:00 PM

I have a variety of deep disagreements with NZ's MMP system (to be explained fully in forthcoming book) which it's pointless getting into here, but I have a specfiic line of questioning that might help you refine your (basically very nice) submission:

1. Why do you take major party overhangs to be more damaging than small party overhangs? After all the large party cases are going to be relatively small distortions - e.g., Labour gets 43/41 of the seats it deserves in one of your examples, whereas the Maori party gets 4/3 of the seats it deserves in the current parliament and could easily have ended up with 7/3. It's almost impossible for large parties to produce such scandalous distortions, ergo they're less of a problem (if proportionality is what you care about), which is the opposite of what you suggest. In sum, as written your submission just seems biased in favor of small parties at the expense of proportionality, which you don't want.

2. You say, correctly, that shrinking the size of parliament makes overhangs more likely as well as more potent when they do occur. But you don't mention that small party overhangs will be most explosively benefitted. Increasing the maori party's overhangs by one (which is what reducing to 100 MPs would generally mean) is a good deal for them for which they'd gladly trade the odd overhang for someone else (In the limit, the Maori Party would do best under straight FPP independently of whether that would help Labour and National.) So even though you appear to be a small-party advocate of some sort you don't follow through fully.

3. Assuming the Maori Party is here to stay, the maori seats almost certainly curse us with overhangs from here on out. Since you don't advocate the elimination of the maori seats you evidently think that proportionality isn't *everything*. But then that leaves you open to the charge of inconsistency. For example, *something* is gained by having a greater proportion of Paliament accountable to a local consitutency and fewer pure party creatures in the House overall. You, of course, don't think that that sort of localized representation counts for much - you don't believe in trading even small amounts of proportionality to get it and so dont even mention it (but of course it's what's guided choices of 75% electorate seats in many MMP settings including in the largest German state (Land) North-Rhine Westphalia). But this is just to say you appear to genuflect before proportionality when it suits you but not otherwise.

One last thing. Your point 10 is *all over the map* and must be deleted in its current form. At the very least you need to properly logically disentangle MMP and (arguable) reforming the executive issues from size of House issues. But really there are just huge numbers of issues unhelpfully raised there which raise objections you can't begin to answer: e.g., more eyeballs perhaps but more of them are those of party drones. Yikes (and remember all the overhangs help with the size of house issues. haw haw!)

Hope that helps... - s.g.

Posted by Anonymous : 4/20/2006 01:43:00 AM

I would have a look at Arend Lijphart's Electoral Systems and Party Systems which scientifically proves that the larger the legislation body, the less disproportionate the electoral results.
Furthermore, look at legislative bodies of countries with a similar population to NZ. NZ is already on the small side.
Parliament should be increased to 161 MPs (160 being the cubic root of the NZ population but needing an odd number to prevent votes being tied).
Also, the number of list MPs should equal the number of electorate MPs.

Posted by Anonymous : 4/20/2006 02:01:00 PM

Fred: I'm not sure if your remarks are directed specifically at me, but, yes, of course, the larger the number of MPs the closer to proportionality you are guaranteed- one can also get that result flexibly by incorporating an Ausgleich seat mechanism - such a mechanism (using the standard rule) would compensate for the Maori party's overhang seat by expanding the parliamentary pie to 186 seats.

As you say, if proportionality is what you care about then more than 120 MPs and reducing to about 50% electorate seats (of the pre-ausgleiched pie) as a starter is where you want to be.

But I was trying to help out I/S who's trying to say that we can (and do) live with small-party overhangs that seriously distort propotionality but we couldn't possibly live with large party overhangs that would much less seriously distort propotionality. That's a hard row to hoe if you worship proportionality (as i/s does, e.g., to the extent of being oppsed to any thresholds for list seats). And then that raises other questions..... hence my earlier post.

hope that helps --s.g.

Posted by Anonymous : 4/20/2006 05:50:00 PM

One option to reduce the number of overhangs (and simplify voting) would be to get rid of vote splitting. Each voter would simply be deemed to have cast their party vote for the party of their chosen electorate candidate.

Posted by Rich : 4/20/2006 07:53:00 PM

Stephen - it wasn't directly at you, but hey

I would like to point out that the 2005 election is the most proportionate election result ever in New Zealand (using the Gallagher Index. The numbers are:
1946-1993 average|11.10%
Which shows that at the moment an overhang doesn't really make a great deal of difference. What makes the difference is when parties get close to the five percent threshold but don't make it over (at each MMP that has been going down).

But the question of large party overhang to small party is interesting. I would say that while the minor party gets a more significant boast out of having an overhang, not just in terms of numbers but also money, it is better for them to have one than the larger parties. A minor party having an overhang is more likely to increase the diversity of Parliament and more likely to bring new ideas in, thus making a difference. A major party overhang MP is more likely to just tow/toe (?) the line.

Posted by Anonymous : 4/20/2006 07:55:00 PM

Stephen: I think it helps if you see this as a submission aimed at defending the status quo against this specific change, rather than a plan for a perfect electoral system. The latter will be beyond the scope of the committee's discussions, so there is no point in my talking about it.

I think overhangs are bad, period - but they're the price you pay for a mixed-member system. If we want to stop them absolutely, then we need to ditch electorates entirely. That's not going to happen (and I'm not entirely sure its a good idea anyway), so my concern is instead to stop there from being any more of them than at present. As for the Maori seats, firstly I think the problem is due to there being electorates, rather than specifically Maori ones, and secondly, I expect that particular point to be done to death by other submitters.

You're probably right on paragrah 10, and it needs expansion into some general points in favour of a larger rather than smaller House. I do think we've seen better work from committees since 1996, and is basically due to the fact that there are more backbenchers to serve on committees, and therefore more person-hours to be devoted to considering legislation. While party-hackery can win out, its still far less of a factor than it was in the pre-MMP days, when Parliament was just a rubberstamp for Cabinet's ruling clique.

Posted by Idiot/Savant : 4/20/2006 09:59:00 PM

Rich: good point about cost (I'll just steal that, if you don't mind), and about changes to the electoral system. There's a reason parts of it are entrenched, and arguably that protection hasn't gone far enough (while at the same time potentially causing long-term problems with seat readjustment...)

Fred: Unfortunately I'd have to interloan it from somewhere. I presume Lijphart is talking about mixed-member systems and not purely electorate ones?

Stephen: I should also point out that its more that we can live with infrequent small overhangs far more than we can regular (50%+ of the time) larger ones.

Posted by Idiot/Savant : 4/20/2006 10:22:00 PM

Thanks to Fred and I/S for your replies. Let me have another go! And Rich, I agree that it's the two votes that have to go and have the book done to prove it...

Fred: The Gallagher Index is an interesting and important measure, but mainly to Poli. scientists. More relevant to practical politics is how the parties that make it into parliament fare. And for *that* the measure that's specific to our MMP system is needed: compare percentages of seats to the adjusted percentages of the party vote (i.e., that result from discarding parties that don't meet the 5% or an electorate win threshold). It's this approprpriately tweaked # that'smuch smaller in pre-2005 MMP elections and that any overhangs blow up and that small-party overhangs *much* more severely distort. Of course there's something stupid and artfiicial about summing the (squared) overhang differences with the other (squared) differences - the latter are technically random or statistical errors in the House-divvying up step of your electoral system whereas the overhang diff (at least from a proportionality-favoring perspective) is a systematic error. Any Gallagher-type statistic just hides that important distinction in my view.

I/S: It's true that the maori seats' problem is a general one: any small party tailored just to win in certain electorates will cause the same sort of problem. (A party built on getting Auckland more representation in parliament could form and target just the 17 Auckland electorates and not contest the party vote much.)
The Maori seats *invite* an instance of that general problem, and given the Maori party, that invitation has been accepted! We've now got one of those nasty problem party + seat combinations - it's literally overhangs every time from here on out. We were very lucky to get away with an overhang of only one in 2005 (if 1400 maori party voters, seeing that they're likely already on an overhang and not wanting to waste their party vote, party vote Green, then bingo the overhang's 2. And if the electorate votes had tilted just a little more to Maori, the overhang easily jumps to 4 or 5.).

That's the future for NZ under 120 MPs: every time, substantial, probably balance of power-holding overhangs, (with the incentives set up to always to try to increase that position).

When you do the 100 MP simulations properly, by way of contrast, the large party overhangs are quite hard to get - only Labour is a threat and if you subtract out most of the maori seats for them in 1999 and 2002 then it's not close even in those cases. And the large party overhangs are much less distorting when they do occur.

So... if you accept the maori seat realities then proportionality is due to take a beating in any case. The marginal impact of large party overhangs caused by 100 MPs is negligible by comparison. It therefore is wrong to say that the "sky will fall" if we have only 100 MPs, it wouldn't: we'll deal with worse under 120 MPs in any case.

The argument form you need therefore seems to be a "straw that breaks the camel's back" argument, i.e., that the minor indignity of a few large party overhangs is just the final straw, after the trashing that proportionality is due to get anyway under 120 MPs etc from small parrties. (I've focussed on the Maori Party case, but it's not completely beyond the pale to imagine Messrs Dunne, Hide and Anderton keeping their electurate strongholds while their party votes wither away completely - small party overhang-arama. Or more likely the Maori Party example will lead to more deals being done trading local and electorate votes with large parties in all these other cases too.)

And you need to make *that* more complicated argument without sounding like you're just biased in favor of small parties and *their* overhangs and distortions. That's not easy to do I think (but you'll think of something, I'm sure!)

And on the pool of talent in parliament issues: the public rightly or wrongly thinks that at least the bottom 20% of MPs are dead wood - that they're useless on select committees, useless at monitoring the govt., will never be in cabinet let alone a major player in cabinet, and so on. So they're no loss at all. An abstract point to the effect that "more is better" (a counterpart of which would hold in every workspace) is weak against this sort of skepticism, isn't it?
Suggestion: develop a "yes the bottom 20% are crap but this proposal would throw out some of the good with the bad, and we need the eggs..." line of reasoning? At any rate, once again, it seems to me, only a much more nuanced version of your thought stands a chance of working and of being persuasive (except with that bottom 20% - those drongos go along with anything. Haw haw!).

OK, whew,... sorry if I've beaten this into the ground! I don't think I'm asking for you to solve every problem with MMP. Rather, I just bear the bad/good news that it's harder to make a cogent argument against the 100 MP proposal than you've allowed, and that is because it's marginal impacts given other MMP infirmities that are at issue.

Posted by Anonymous : 4/21/2006 04:41:00 AM

Large party overhangs would be hilarious (there were even brief thoughts that these might affect Labour in 2002).

Think about it - no Michael Cullen, no Speaker Wilson, no Don Brash.

Posted by Graeme Edgeler : 4/21/2006 10:06:00 AM

"Fred: Unfortunately I'd have to interloan it from somewhere. I presume Lijphart is talking about mixed-member systems and not purely electorate ones?"
->it's a comparison of all electoral systems, a rather short $100 book. It is, however a fantastic book

Posted by Anonymous : 4/21/2006 10:07:00 AM

Graeme: I agree, it would be funny - but hardly desirable.

Posted by Idiot/Savant : 4/21/2006 10:17:00 AM

Following up on Graeme's point... yes, large party overhangs are perfectly possible given 120 MPs: in 2002 Labour's party vote needed to go down to below about 38% (while keeping the 45 electorate seats) to trigger it. I/S's Point 3 *seems* to suggest that winning more than 70% of the electorates ("which has never happened" or whatever) would be required, which is false, so it too should be deleted in its current form.

Graeme: I hadn't thought about the fact that the possibility of large party overhangs does put people in jeopardy that normally aren't. Two cheers for large party overhangs then! I personally hate the feeling that there's essentially no way to stop, say, Margaret Wilson from getting back into parliament: "We can't stop her, nobody can..." is a horrible, bete noire.... Overhangs help a little (but only a little) with that and insofar as 100 MPs boosts the likelihood of large party overhangs, to that extent I think it helps with one of MMP's problems. A small boost for 100 MPs to be sure (particularly since the small party overhang future we have to forward to makes large party overhangs harder to get) but it's real nonethless.... Wow, yes, if National just does a little better electorate-wise (25 rather than 21) then given 100 MPs in 2002, no Don Brash!

Posted by Anonymous : 4/21/2006 01:58:00 PM

Stephen, the way to deal to that problem is to go to an open list, in which voters can re-rank the list. This can be done by giving each person a vote on the list (for the party they support), and if that person crosses a threshold (five percent prehaps) then they move to the top of the list. If multiple people get above the threshold, then the person with the most votes go to the top. If people decide not to rank someone, then it is an endorsement on the current ranking.

Of course, nothing has changed since FPP. My vote in Wellington Central couldn't decide whether or not Winston was an MP. It's just a different way of looking at that problem

Posted by Frederick Aloysius Weld : 4/21/2006 03:23:00 PM

Frederick: Yeah open lists do help a bit with this problem, you're right (at leat as a kind of theoretical point). That party lists are "closed" has the advantage that voters can at least broadly calculate what personnel different levels of popular support for a particular party will elect, hence what (given what the polls say) their party vote will probably lists jeopardizes that calculability to some extent so there's a trade-off. Other things being equal, however, that 100 MPs would do something (albeit small) to ping elector-proof parliamentary sinecures sounds good to me. [If you don't hold other things equal, then electoral system discussions dissolve, e.g., "You say that 100 MPs shrinks the pool of talent to be in cabinet... well, that just goes to show we need to allow the PM to appoint non-MPs to the executive the way they do in Denmark (or wherever it is)..." And so on and on.]

Posted by Anonymous : 4/21/2006 04:55:00 PM