Save MMP!

What Germany tried

Germany’s three fixes

Germany met the overhang a generation before New Zealand. Between 2013 and 2023 it tried three fixes. Each solved the problem it aimed at and created the next one.

In 2009 Germany stood where New Zealand stands now. It has tried every remedy but one, and we can see how each turned out. We do not have to repeat them.

The problem Germany had

Germany has used MMP since 1949 (see the history). For forty years an overhang was rare. After 1990 it became a regular feature. The two large parties lost votes to newer parties and still won most electorates, so their electorate seats ran ahead of their party vote.

Until 2013 Germany did what New Zealand does today. A party kept its extra seats, and nothing was done to restore the balance. The extra seats never reversed an election, but they shifted the balance more than once:

  • 1994: 16 overhang seats. The government’s majority was 10 seats where the party vote had earned it 2.
  • 2002: 5 overhang seats. They made one party the largest in parliament, by 251 seats to 248, when the two large parties had the same share of the vote.
  • 2009: 24 overhang seats, all on the government’s side. A close election would have been decided by them.

In 2012 Germany’s constitutional court set a limit. An overhang with no correction is acceptable only up to about 15 seats, in a parliament of about 600. Beyond that it distorts the proportional character of the election. A first attempt at new rules, passed in 2011, was struck down in the same ruling before any election had used it.

The three fixes in numbers

ElectionRulePlanned sizeOverhang seatsSeats added to balanceMembers
2009Tolerate598240622
2013First fix598429631
2017First fix5984665709
2021Second fix59834104736
2025Third fix630nonenone630

Source: Deutscher Bundestag, Datenhandbuch; the Federal Returning Officer for 2025.

The first fix, 2013: add seats to balance it out

What it did. Every party with an overhang kept its extra seats. Parliament then grew, with extra list seats for the other parties, until every party’s share of the seats matched its share of the party vote.

What worked. Parliament was proportional again. No party gained votes in parliament from an overhang. All four established parties agreed to the law.

What went wrong. Parliament kept growing: 631 members, then 709, then 736, against a planned 598. The rule multiplies. In 2021 one small regional party had 11 overhang seats, and most of the 104 balancing seats were needed to make up for those 11.

For New Zealand: on 2026 election-night figures this fix would have meant at least 146 MPs with the Māori electorates set apart, and with them counted in, no workable size at all.

The second fix, 2020: a partial brake

What it did. Up to three overhang seats were left uncorrected, and a party’s overhang in one state could be set against its list seats in another. A cut in the number of electorates was planned and later dropped.

What worked. Very little. The parliament elected in 2021 was the largest Germany has had.

What went wrong. The lasting answer on offer was fewer electorates, and that would have cost the governing parties’ own MPs their seats. They could not agree on it. The law passed against the votes of the opposition.

For New Zealand: the parties that gain from a rule find it hard to change that rule. So the public has to ask for the change.

The third fix, 2023: turn winners away

What it did. Parliament was fixed at 630 seats, with no overhang and no balancing seats. In each of the sixteen states, a party’s electorate winners are ranked by their share of the local vote. They take their seats only as far as the party’s share of the party vote in that state covers them. The winners with the weakest results miss out. This is the German rule.

What worked. Parliament is proportional and its size is fixed. In 2024 the constitutional court upheld the rule unanimously.

What went wrong. At the 2025 election 23 candidates who won their electorates did not become MPs: 15 from the CDU, 4 from the AfD, 3 from the CSU and 1 from the SPD. The law passed against the votes of the opposition, and the government formed in 2025 promised to seat every winner again. It has not yet agreed how.

Who lost their MP. The winners who miss out are those in the closest contests, and most of those are in cities. Among the 23 are both Frankfurt electorates, Munich South, Nuremberg North, Augsburg, Heidelberg, Mannheim, Mainz, Leipzig I, Halle and Bremen I. In 19 of them another local candidate still reached parliament through a party list. In four, nobody did. Darmstadt, a city south of Frankfurt, has no MP from any party. Nor has the university town of Tübingen, nor Stuttgart II, one of the two electorates of Stuttgart, nor Lörrach–Müllheim on the Swiss border.

For New Zealand: picture Auckland Central, Wellington Bays or Christchurch East voting, choosing an MP, and being told that the winner will not sit and that nobody will represent the electorate in Parliament. We think this price is too high. An electorate should keep the MP it chose.

Germany no longer tolerates an overhang. We tolerate it, and take away every reason to create one.

What we take from it. Our fix uses the same test as the German rule: what a party holds must be covered by its party vote. Germany applies the test to the seat, so the winner does not become an MP. We apply it to the vote in the House, so the winner becomes an MP and only the vote in the House goes.

Germany is now looking for the same thing. The parties that formed the government in 2025 asked a commission to find a way for every electorate winner to enter parliament, while parliament stays proportional and about the size it is now. Our fix would be one answer. On the 2025 result all 23 winners would sit, and parliament would still cast 630 votes, shared out by the party vote. It would have 653 members, so the size holds in votes and not in members. German law also expects all MPs to have equal rights, and a non-voting MP has not been tested there. We do not know whether the commission has looked at this idea.

What Germany’s experience shows

  • The harm is in the extra votes, not the extra members. From 2013 to 2021 Germany’s overhang MPs still sat, and no party gained votes in parliament from them. The complaint in those years was the size of parliament.
  • Each fix moved the cost somewhere else. First other parties’ voters paid for it. Then the taxpayer paid, for a parliament that kept growing. Now electorates pay, by losing the MP they chose.
  • A fix passed by one side does not last. The laws of 2011, 2020 and 2023 were each passed by the government against the opposition, and each was challenged or reversed. Germany’s rules for its next election are still uncertain.

How our fix differs

Our proposal takes one thing from each German fix and leaves the cost behind.

From the first fix

No party gains votes from an overhang. But no seats are added, so Parliament does not grow to balance it.

From the third fix

The party vote alone decides each party’s votes in the House. But every electorate winner sits, so no electorate loses its MP.

Our fix

Every electorate keeps the MP it chose. Parliament keeps 120 votes. An overhang MP serves the electorate and does not vote in the House.

We tolerate an overhang. We stop it from distorting proportionality.

See Germany and New Zealand compared, and the German states and New Zealand compared.

And because a fix passed by one side does not last, we want ours passed with broad support and then entrenched, which takes three-quarters of all MPs.

See the fix