Who needs thresholds, anyway?

There have been no surprises in New Zealand since my earlier report. As I suggested, late counting delivered an extra seat to Labour, giving them 45 to the National Party’s 58. Two other parties, Winston Peters’s New Zealand First and the Greens, cleared the 5% threshold, winning nine and seven seats respectively, and the free-market ACT held its single electorate seat.

Postal votes are yet to be counted, but they are unlikely to change the picture. The Māori Party was eliminated from parliament, losing both its electorate seats to Labour. United Future also disappeared, with the retirement of its sole MP.

National’s 46.0% of the vote is a drop of just 1.0% from 2014, which is pretty good for a third term government. Labour’s increase was much more dramatic, improving by 10.7% to 35.8%, but most of it came at the expense of minor parties – particularly the Greens, whose 5.9% was down 4.8% from last time.

Peters now has a theoretical discretion as to whether he delivers a majority to National or to the Labour/Greens combination, and he is revelling in his status as kingmaker. (He’s that sort of person.) But as I pointed out before, it would be politically impossible for him not to back National, given that they lead Labour by more than ten percentage points.

Long-time readers will be aware that I rather like the New Zealand electoral system, but it has two notable drawbacks: the rather high threshold for representation, and the capricious exemption from it for winning even a single electorate seat. (Germany shares the first but not the second.)

Neither feature, however, really mattered yesterday. Only ACT won an electorate seat from below 5%, and its party vote was so low (0.5%) that it was never going to bring in any list MPs. The largest unrepresented party is the centrist newcomer TOP (“The Opportunities Party”) with 2.2%. If there had been no threshold, TOP would have won three seats and the Māori Party one, with the seats coming two each from Labour and National.

That wouldn’t really change anything, although it would make the already difficult task of forming a Labour government just a bit more difficult. But in other circumstances, the threshold can wreak havoc. In 2014 the Conservative Party came fifth in terms of votes, but with 4.0% it won no seats; three parties with fewer votes – Māori, ACT and United Future – were all represented.

This year the surge in Labour’s support carried the danger that, because so much of it was coming from the Greens, it might have pushed them below the threshold. It’s impossible to say, but some of the faltering in Labour’s momentum in the last week could possibly have been due to an awareness that it would do more harm than good if it resulted in its ally dropping out of parliament.

Germany can provide equally striking examples. In the 2013 election, the three parties of the left won a majority of seats, despite the fact that their aggregate vote had gone down, simply because two right-of-centre parties (the FDP and AfD) just fell short of the 5% mark.

The desire to avoid a proliferation of tiny parties in parliament is understandable, but it’s not worth the arbitrary nature of the current setup. A 3% threshold, as in Greece, would be much less likely to distort the outcome.

In Germany, of course, it was not just small parties in the abstract that were the problem; the threshold was designed specifically to prevent representation of the far right. The NPD fell short with 3.6% at its peak in 1969, and AfD managed 4.7% in 2013. But with AfD now polling well clear of 5%, that purpose has become an irrelevance.


3 thoughts on “Who needs thresholds, anyway?

  1. German-style all-or-nothing thresholds have three downsides:

    1. I am mystified that German electoral law doesn’t allow alliances of parties within the same State-wide constituency, or nationally. The CDU list in Thuringia can link with the CDU list in Saarland, but not with the FDP list in Thuringia. Surely joinder of cross-party lists would be an excellent way to encourage parties to form pre-election coalitions, rather than announcing them Winston Peters-style after voting has closed. Not only would it help win an extra seat here and there by combining surplus votes (as in Israel), but it would save the FDP from dropping below the threshold and disappearing from the Bundestag for the next four years. And inasmuch as it would mean, say, a CDU supporter could vote their first preference for the CDU without worrying that they might need to lend a vote to the FDP to keep it above water, it would encourage sincere voting and widen voter choice.

    (Is there no move for this in New Zealand, at least? Is this my Australian bias showing through?)

    2. Two or more seats – indeed, nearly 40 at a time, in the case of the German Bundestag, currently at 709 seats – can ride on a small number of votes. This is inconsistent with the principle of proportional representation, being more like the US Electoral College or the pre-1948 Australia Senate. Proportionality should mean, ideally, that only one seat flips at a time for any given parcel of votes. (However, those parcels of votes need not be exactly equal, at least not among the smaller party lists — see #3 below).

    3. The German threshold is based on 5% of all valid votes cast. It would make more sense to base it only on 5% of the votes for parties over the threshold – if that sounds circular, what I mean is, eliminate the lowest-polling list unless it has at least one-nineteenth as many votes as the combined total for all higher-polling lists. Under the German rule, as I understand it (unless it’s been amended recently), eg, if two minor-party lists poll (say) 4.8% and 4.9% respectively, they would both be excluded (despite together representing 9.7% of the votes). The remaining 90.3% of votes would determine 100% of the seats, and an absolute majority could be won with just 45.2%, even with nationwide allocation (unlike Poland and Portugal). It is possible that by voting for, say, the Greens you could help raise the threshold, exclude the FDP, and help the CDU-CSU or the SPD win an absolute majority. This is perverse.

    Fortunately it can easily be fixed. Under my proposed tweak, you would first eliminate the lowest party because it has only 48 out of every 1,000 votes still in the equation; however, after it’s eliminated, the lowest party remaining now has 49 out of every 953 votes still being counted, which is 5.14%. The threshold for the 4.9% party would not be affected by the “erased” votes cast for a different party that polled even lower.

    4. I’d go even further. Instead of “bin all votes below the threshold” (even if the threshold is calculated more sensibly, as in #2 above), I’d raise the threshold to 10% but give votes below the threshold a lower weight. The lower parties’ votes should be recalculated by squaring the total and then dividing the result by the threshold. So, eg, 5% of actual votes would be counted as 2.5%. 8% would count as 6.4%. 2% would be converted into 0.4% and unlikely to win any seats except in a huge assembly. (Caveat, to be pedantic – the 5% party’s votes would count, not as 5/100, but as 5/95 of the total; a 7% party’s adjusted total would be not 4.9/ 100 but 4.9/ 97.9 of the total, and so forth, but you get the basic idea).

    This way, there would still be an icnrntive to vote for your first-choice party; your vote won’t be completely binned (unless it’s polling below 2-3%); your vote can help that party win seats. It just wins them on a sliding scale if it’s below the threshold.


    1. Thanks Tom. I think the ability of thresholds to shift large numbers of seats for a tiny change in votes – especially if you have a relatively large threshold and a large number of MPs – is a serious problem, and I don’t think there’s a good way around it. You could introduce some sort of sliding scale, as you suggest, but (a) it would be difficult to win public understanding and acceptance for and (b) it would make explicit the fact that the threshold system involves making some people’s votes worth more than others. I’m happy just scrapping thresholds altogether, but I understand some people will be concerned about the fact that in a large parliament a party might get someone up with just a couple of tenths of a per cent.

      I also quite like the Israeli system of combining surplus votes; it’s sort of a distant cousin of the old group voting tickets in Australia, but much less objectionable. But it only applies to parties that have already crossed the threshold; if you’re below the threshold your votes aren’t counted regardless of the surplus agreements.


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 )

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.