The NZ electoral system referendum

Results have now been released from the 26 November referendum on the New Zealand electoral system. The referendum consisted of two parts. In Part A, voters were asked “Should New Zealand keep the Mixed Member Proportional (MMP) voting system?” In Part B, voters were asked “If New Zealand were to change to another voting system, which voting system would you choose?”, with a menu of four choices for the change option: First Past the Post, Preferential Voting, Single Transferable Vote, and Supplementary Member.

In Part A, 57.77% voted to keep MMP. This, of course, renders Part B moot. Nonetheless, comparing across the two parts is interesting. While 2,194,774 valid votes were cast in Part A, there were only 1,509,157 in Part B. Informal votes, defined as “when the voter has not clearly indicated the option for which they wish to vote” were just 62,469 in Part A, but a whopping 748,086 in Part B.

If we can assume that essentially all of the 926,819 voters who voted to change from MMP would have indicated a preference over the four change options, then it would seem that approximately half of the “keep MMP” voters did not bother to select an alternative system.* One wonders how many of these voters simply were confident MMP would be kept, so it was not worth taking a stand on (and bothering to learn about) the alternatives, and how many did not understand that they could choose an alternative even while voting for the current system.

Rationally, any MMP supporter should prefer STV over all the others offered, given that STV is the only proportional system among the alternatives.** Yet STV obtained only 252,503 votes (16.7% of valid votes in Part B), placing third among the alternatives. If just under two thirds of the informal vote in Part B had been cast by MMP voters indicating STV as their fallback, then STV would have beaten out FPTP in Part B.

A referendum in 2014 between the two proportional systems would have been interesting, had “change” won in Part A. Instead, however, the old FPTP system dominated Part B. It was probably expectation of precisely such a result that generated such indifference over Part B, as few could imagine FPTP ever beating MMP in a second referendum were one to have been triggered.

The results in Part B, with percentages based on valid votes only, were:

    48.66% FPTP
    24.14% SM (i.e. MMM)
    16.72% STV
    12.47% PV (AV)


* 685,617 is the difference between the valid votes cast in the two parts; this is equivalent to 54% of the 1,267,955 votes for keeping MMP.

** Unless what someone really likes about MMP is the single-seat districts, yet would not want FPTP again. I suppose such a voter might rank Preferential Voting–that is, the Alternative Vote–as second best. And some could argue for “SM” (MMM) as being most similar to MMP, though it is far worse than any option other than FPTP for supporters of parties other than the big two.

30 thoughts on “The NZ electoral system referendum

  1. The electorate-level i.e. constituency-by-constituency results of the voting system referendum show an extremely strong correlation coefficient (0.97) between the percentage of informal votes in Part B and the “Keep MMP” share of the total vote (that is, including informal votes) in Part A. Moreover, the correlation coefficient between the “Change” percentages in Part A and the joint share for FPTP and SM, as a percentage of the total number of votes (valid and informal) cast in Part B is also 0.97.

    However, the correlation coefficient between “Keep MMP” and STV is extremely weak at 0.07, but relatively strong with respect to PV/AV, at 0.59.

    These figures suggest that the bulk of informal votes in Part B were indeed cast by MMP backers who for one reason or another ignored the second part of the referendum, while most of the voters who wanted change preferred either FPTP or SM. As for STV, its low correlation coefficient with respect to “Keep MMP” may be indicative that those who backed that option were a mixed bag of some voters who wanted change but nonetheless wished to retain a proportional voting system, and others who wanted to retain MMP but nonetheless chose STV as a fallback option. That said, it would appear that a sizable number of MMP voters in Part A went instead for PV/AV in Part B of the referendum.

  2. As a data-point, I would prefer STV to MMP. However, knowing that MMP would be against FPTP in any second referendum, I voted ‘No Change’ and STV.
    A ‘rank these 5 systems in order of preference’ poll would have been interesting, as would the argument about which method to use to resolve the vote.

  3. “As a data-point, I would prefer STV to MMP”

    I’m almost exactly on the opposite side of the world from New Zealand, but we think alike. At the time of the initial switch from FPTP (though I prefer calling it SMP) to MMP, I though that it was too bad they didn’t switch to STV, but that MMP was an improvement.

    My reaction was based essentially on the similarity between New Zealand and Ireland in size, in the historical heritage of using the Westminster model, in the historical heritage of using single member districts, and in the presence of a Visible Minority, which was a large factor behind the adoption of STV in Ireland in the first place, though in both New Zealand and Ireland alternative methods were used to deal with the issue. And STV turned out to have a decent record in Ireland, though I don’t think MMP has gone badly on the whole in New Zealand.

    I haven’t seen any reason to revise my opinion from twenty years ago.

    • The Ireland-New Zealand comparison Ed offers is interesting, and I have thought STV would be a reasonably good choice for NZ, too. (Had I a vote, I would have voted to keep MMP, however, and likely even if I had thought STV was a reasonably live option.)

      One possible argument against STV in New Zealand is that there seems to be nothing in that country comparable to the intense localism or what is sometimes called “friends and family” approach to politics of Ireland. Whether such a culture would develop if NZ adopted STV, and whether it would be beneficial, is something I’ll leave to others to debate.

      There is little doubt that MMP has increased the centralization of parties and the nationalization of politics in New Zealand (even though it was hardly “localist” before). Again, whether that is good or bad for the country’s politics overall is something I should leave to others.

      If, however, I view the question through my Green-tinted glasses, there really is little question that MMP is preferable to STV.

  4. “The greatest support to keep MMP came from Maori voters, while the strongest current for change flowed from staunch National seats, a breakdown of voting by electorates shows.” (NZ Herald, 13 Dec.)

    No surprise there, of course.

  5. Does STV universally cause increased “localism of what is sometimes called friends and family”? I heard that it causes an increase in pork barrel spending because MPs compete against one another even from with in the same party.

    I agree that MMP is a better system than STV because it is easier to understand, and one does not have to deal with the distribution of surplus votes. It is also more proportionate.

    The choice of an electoral system for a country should met that countries needs and be the right fit.

    In NZ, MMP was considered over STV because of two elections in the late 70’s and early 80’s lead to a reverse plurality. STV can produce results like that, look at Malta which ironically produce a perverse reverse majority result in 1981, and NZ second reverse plurality.

    When ever a country has a perverse result, the electoral system is changed, any examples of a country with a perverse election result, but the system was not changed?

    “5. When the parties are fairly evenly matched in voter support, then the process of translating votes into seats can produce unexpected and distorting outcomes, as was demonstrated in a spectacular fashion by the “perverse” result of the 1981 election. In that year the MLP gained 49.1% of the popular vote yet obtained 52.3% of the seats in parliament and formed the government. While some disproportionality between vote and seat percentages has been quite common, what made the situation unprecedented was that the rival party (the PN) had actually obtained a majority of the popular vote but was not rewarded with a majority of the legislative seats. The same situation occurred again in 1987 and once more in 1996, but by then a newly adopted constitutional amendment provided that the party with a majority of the popular vote would be awarded a sufficient number of additional seats to give it a legislative majority.” from

    STV would be the perfect electoral system for local government because parties are less important at that level of government. Funny that Scotland and NZ use this system for it’s local government.

  6. Talking to a bunch of MMP supporters, quite a few gave the tick to FPP in part B because they believed that FPP was much less likely to beat MMP in a second referendum than the other options such as SM (the preferred choice of various National MPs and the anti-MMP group Vote for Change).

    A whole bunch of others refused to answer Part B because they didn’t wish to endorse any alternative system with their vote so left it blank. It felt like nearly every MMP supporter I knew fell into one of these two camps.

    There was a lot of confusion about it though. I found myself correcting a lot of people on Facebook about being able to vote in Part B if you voted to retain MMP et cetera. The host of TV3’s election show John Campbell made that mistake when reading the preliminary results (he said that only voters who voted for change could vote in Part B) and then complained that no one understood it.

  7. First, as to why MMP supporters failed to tick STV as their favourite alternate system, I believe the Campaign for MMP considered advising supporters to do so, but feared this would be a diversion from the focus of their campaign. Second, Suaprazzodi asks for examples of a country with a perverse election result where the system was not changed. In Canada in 1979 Pierre Trudeau’s Liberals got 40.1% of the vote, but only 114 MPs. Joe Clark’s PCs got only 35.9% of the vote, yet elected 136 MPs and formed the government with support from six Créditiste MPs, giving them a one-seat majority. In the 1980 Speech from the Throne, the Trudeau government promised to appoint a committee to study the electoral system; however, none was ever struck. Opposition to even modest reforms among MPs was intense. As usual, the party’s regional strongholds elect most of its MPs, who fail to act on their sympathies with unrepresented supporters elsewhere. A vicious circle, so the parade of stronghold politics continues.

    • Actually, there are many more cases of “perverse” results with no resulting reform than cases with reform. Of the latter, in recent decades, there is only once case (NZ). I wrote a chapter about this:

      Matthew Soberg Shugart, “Inherent and Contingent Factors in Reform Initiation in Plurality Systems,” in To Keep or Change First Past the Post, ed. By Andre? Blais. Oxford: Oxford University Press, 2008.

  8. From what I saw in New Zealand over the election and referendum weekend, it sounded like many MMP supporters deliberately voted for FPP because they thought it was the most unattractive option to New Zealanders should MMP not have won in Part A. The issue of voter confusion over whether they were allowed to vote in Part B if they supported MMP in Part A is interesting and not surprising considering how poor the public education campaign was.

    Yes, I heard that STV has never really taken off in NZ, despite the fact that local councils are allowed to use it (and few do). I agree with Matthew’s comments about how centralised the party (and political) system is in NZ. The parties (apart, perhaps, from the Greens) allow very little participation by rank-and-file members when it comes to policy and leadership choices. What we have in Scotland is (perhaps surprisingly) more democratic (if you equate democracy with letting the ‘grass roots’ have a lot of say). Very interesting.

    Thanks for these comments, everyone.

    • I can understand the logic of an MMP supporter voting for the least likely of the alternatives to win a potential runoff against MMP. And that certainly would be FPTP. But I would always caution that one should “do no harm”, and that such a vote could always come back to bite. Three years can be a long time in politics; what if the concurrent parliamentary election had produced some outcome that proved politically detrimental to proportionality and coalition government (such as Winston Peters holding the balance of power, or Greens pivotal but divided over providing confidence and supply)?

      The concern of insincere voting in one election in order to increase the odds of victory for one’s preferred option in another arises regularly from “open” primaries in US states. Do some, for example, Democrats vote in the Republican primary, to try to tip the other party’s nomination towards an unelectable candidate. I think most studies have found that such behavior is actually rare. If this sort of insincere voting for FPTP really was widespread in New Zealand it would be quite interesting for political science.

      Are the original ballots for the referendum, or images of them, kept so that researchers can actually investigate this possibility?

  9. “The parties (apart, perhaps, from the Greens) allow very little participation by rank-and-file members when it comes to policy and leadership choices. ”

    That is very interesting. One of the reasons the electoral system in New Zealand changed in the first place is that voters in the early 1990s were increasingly fed up with first a Labor government and then a National government pursuing neoliberal policies which were unpopular with those parties’ own supporters, let alone the wider electorate. If the grass roots had had a stronger role in those parties, maybe there would have been a grass roots revolt within the parties instead of a splintering of the party system and change to the electoral system.

    I think the degree of openness of political parties within a polity to change pushed from below, including the ability for ordinary party members without connections to the elite to realistically seek nominations for offices, is a crucially important feature of political systems and should be getting more attention from political scientists.

  10. I wonder if Tom Lundberg would agree that the reason Scots voters are more likely to want the ‘grass roots’ to have a lot of say may relate to geography? Scotland has a variety of regions and regional institutions. Little New Zealand, about four million people in the middle of the ocean, seems to have no discernable regional feelings.

    To a Canadian, the fact that the South Island has only 24% of the population, and a declining share, should generate regional grievances making a national list system impossible to imagine. But Kiwis don’t seem to care. The Labour Party holds six regional nomination conventions for their list candidates, then folds the six lists into one with no reported controversy. Their 16 regional administrative divisions and 67 territorial authorities seem to leave them cold too.

    • Wilf’s characterization of New Zealand’s relative lack of regionalism seems about right to me. Except…

      In the Green Party, the one with the most grassroots involvement in candidate selection, six of its fourteen MPs in the new parliament are South Islanders. That’s significant over-representation.

      Whether that is due to pent-up regional sentiments in New Zealand that only the Greens have the party rules to represent, or to Green sentiment simply being more prevalent on the South Island, relative to population, I can’t say.

      (Greens use a member ballot, via STV, to set the list, subject to some criteria on minimal representation of women and–yes–South Islanders. Both women and South Islanders received sufficient votes from the membership to be represented beyond these minimal quotas. Eight of the fourteen, including three of the South Islanders, are women.)

  11. The Kiwiblog analysis is really interesting.

    The numbers on South/North balance in the Green Party caucus are very slightly different from mine. I suppose I could have just made an error on one of the MPs, but I wonder if our criteria are different.

    David Farrar (Kiwiblogger) does not say how he defines someone’s island. I used the island of the electorate in which the candidate was nominated. Regardless of where someone is from originally (which, I will admit, is the criterion I normally use in my research), in a system with districts as well as a national list, it seems sensible to locate a politician’s base by the district in which he or she runs.

    (It is interesting that the Greens, a party that seeks only the nationwide party-list vote, and has only once won an electorate, nonetheless is organized on the basis of the electorates. Best of both worlds…)

  12. Thanks to Wilf Day for the comment above about regionalism in Scotland. I can’t be sure, but it’s very possible that region matters more in Scotland. I’ve heard this in the four+ years I have lived here, but have not seen much research done on it. It’s something to look at during my upcoming research leave. Scotland’s MMP system uses eight electoral regions (and parties use these regions for the list rankings by party members), while NZ, of course, does not use regions at all. Scotland no longer has a regional tier of local government (these were abolished in the 1990s), and the current SNP government is actually planning to centralise public services (like the police). The SNP was also against having a local income tax that would be set by the 32 local authorities themselves (they wanted the rate set centrally (by the Scottish Government), while the Liberal Democrats, who supported local income tax in principle, wanted the local authorities to set the rates; this prevented agreement). I suppose it is not surprising for a nationalist government to want to overcome local and regional ‘issues’ and centralise things on a pan-Scottish basis. This is what we are seeing now, and the SNP’s majority means it can implement its wishes here.

    I agree with the comments about NZ not appearing to have too much in the way of regional sentiment, apart from what was noted about the South Island. There could be some element of urban/rural cleavage (or Auckland against the rest of NZ!), but I don’t think this is too politically significant.


    • Very interesting, Tom. So do all the (significant) parties in Scotland allow members to participate in ranking their lists? Do they use STV, like the NZ Greens? Is membership broad and open?

  13. Good news, you will have to wait a month but (per Elections NZ)

    “Polling place by polling place and split vote information for the Referendum on the Voting System is expected to be available from 1 February 2012.”

  14. Matthew,

    I had a look at my interviews with the leaders of the main (five) parties in Scotland. All allow the members to rank candidates for constituencies and regional lists, and it looks like STV is normally used (I suppose it would be AV in the case of the single-member constituencies), but some parties have mechanisms like ‘twinning’ and ‘zipping’ (Labour) to assist women and ethnic minorities; the Greens also try to ensure equality between male and female candidates on the lists (they don’t normally nominate constituency candidates). The Greens seem to be the most open in terms of who can stand as a candidate (anyone who has been a party member for at least a year), while other parties appear to have some kind of control over who can stand (an application process, and this caused problems for Labour early on because it was controlled from London; the new leader, Johann Lamont, will have more autonomy).

    I’ll need to go over these findings more carefully and get more detailed information, but in general, it’s clear that the parties in Scotland are much more democratic (certainly less elitist and more open) than their counterparts in New Zealand, which took me by surprise. Perhaps there really is something to the aspiration for a ‘new politics’ in Scotland.

    Thanks for the photos – it looks like you’ve had a great time overseas. Best wishes for 2012, Tom

    • Thanks much for that information on Scottish parties, Tom.

      One amendment, however. Drawing from what you say and from what I learned from my interviews in New Zealand, I might say that the parties in Scotland are much more democratic than their counterparts in New Zealand, other than the NZ Greens.

      In fact, the procedures you sketch from Scottish parties seem quite similar to those of the Green Party of Aotearoa New Zealand.

  15. Yes, Matthew. I think Green parties around the world are typically quite democratic in how they operate, though this can cause problems, particularly if they enter government. In New Zealand, Greens have avoided getting too close to the larger parties, which has probably protected them from the kind of difficulties experienced by just about all the other small parties there.

    NZ Greens did very well in the recent election, but their Scottish counterpart did not. I wonder if the 2003 result in Scotland (7 MSPs) was the high water mark (only 2 MSPs in 2007 and 2011) for the Scottish Greens? Our local elections in 2012 (May) will be interesting. There are five Green councillors here in Glasgow, which is not what one would expect for a city with a lot of poverty and social problems. It will be interesting to see what happens in Scotland’s second STV election (you’ll need to set up a thread when the time comes!).


  16. Someone with the NZ Greens whom I interviewed was quite adamant that the democratic nature of the party was both a strength and a weakness. Another was equally forceful in saying that the party’s success to date was based on never having to compromise anything (by entering or offering formalized support to a government), but that this trend could not continue indefinitely. Eventually, the party will need to make a choice on governance, and that would necessitate convening a general meeting of the party delegates. In such a case the weaknesses of the democratic procedures in the party could be on full display.

    When are the Scottish local elections?

  17. No wonder why in NZ they did not tick STV in the second part, the commercial of it is really confusing, especially about the surplus votes. No mention of percentage and quotas.

    This other video shows

    With the STV commercial, they are using a weigh sample of transferring of the votes.

    Are they using a weighed sample or are they going to transfer all the votes at a fractional percentage of whatever the surplus is?

    They were asking people on the street about MMP, and they did not understand it.

    How often does the average person understand electoral systems? Especially coming from a FPTP background, every other electoral system except closed List PR is much more complicated and requires more knowledge to cast one’s vote.

    How do you example to people how does MMP or STV work?

    If there electoral reform in your country, your country uses FPTP, and there is an opportunity in a referendum to change the system to MMP? How would you persuade people on the fence to vote for it?

    • I always say that electoral systems are a bit like cars (or DVRs, or insert your favorite technology example here): You do not have to understand how they work in order to use them effectively.

      Voter behavior, in the aggregate, shows that voters have learned to use MMP sensibly, even if “person on the street” interviews (or even polling) suggest lack of basic understanding of how it works. The same could be said about Germany, or many other countries and electoral systems.

      I have no idea how my car works, but I know where the relevant controls are located and how to make the car go or stop, as needed.

  18. Why should voters choose proportional representation over a plurality/majority system? Why did the referendum in Ontario, Prince Edward Island, and British Columbia fail to change the electoral system?

    For the U.S, the case for proportional representation is to end gerrymandering, and to ensure fair representation of minorities without it being forced with gerrymandering.

    There are lot of interests that need to be represent that are not, because of the single member districts. Most single member districts are artificial not natural communities of interest. Any fair system of redistricting is futile. California’s redistricting commission is a good start, hopefully it will later lead to a move toward PR.

    I do like the car analogy with Proportional Representation. It is amazing that NZ changed it’s electoral system with a referendum in 1993, and it has now kept the system in 2011.

Leave a Reply

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

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

Google photo

You are commenting using your Google 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 )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.