Closer to our size
German states and New Zealand compared
New Zealand has about as many people as one mid-sized German state. Thirteen of Germany’s sixteen states elect their parliaments by MMP, each with slightly different rules. Together they show what happens to an overhang in a parliament of our size.
Why look at the states
Germany as a whole is hard to compare with New Zealand: it has 84 million people and a parliament of 630 (see Germany and New Zealand compared). Its states are a better match. Saxony’s parliament is planned at 120 members, the same as ours, and Hesse’s at 110.
One difference remains. A German state has a constitutional court above its parliament, and the federal government above that. New Zealand’s Parliament has neither.
Six elections at a glance
| State and election | What happened | Size of parliament |
|---|---|---|
| Baden-Württemberg 2026 | One party won 56 of 70 electorates on 29.7 percent of the party vote | 157 members, planned at 120 |
| Schleswig-Holstein 2009 | A government’s majority rested on three overhang seats; the court ordered a new election | 95 members, planned at 69 |
| Saxony-Anhalt 2026 | One party won 38 of 41 electorates on 43.8 percent of the party vote, with no overhang | 83 members, as planned |
| Bavaria 2023 | The two votes are counted together, and there is still an overhang | 203 members, planned at 180 |
| Hesse 2018 | Balancing seats were added; a one-seat majority hung on the count | 137 members, planned at 110 |
| Berlin 2026 | A recount moved one electorate by 34 votes and removed five seats | 153 members, down from 158 |
Four cases that matter for New Zealand
Baden-Württemberg 2026: our near twin
Baden-Württemberg plans for 120 members and has 70 electorates. New Zealand has 120 and 71. In March 2026 the CDU won 56 of the 70 electorates on 29.7 percent of the party vote. Eight months later National won 42 of our 71 electorates on 29.2 percent.
The state balances an overhang by adding seats. Its parliament grew to 157 members: 14 overhang seats and 23 balancing seats for other parties. Proportionality held, at the price of a parliament almost a third larger than the law intends.
For New Zealand: the same balance of electorate and list seats produced a large overhang at almost the same share of the vote. That supports the view that the balance between electorate and list seats is what invites an overhang.
Schleswig-Holstein 2009: a court steps in
In 2009 one party won 34 electorates on about 31.5 percent of the vote. The law balanced only part of the overhang, and three seats were left uncorrected. A government took office with a majority that rested on those three seats, although the opposition parties had won more votes.
In 2010 the state’s constitutional court held the law unconstitutional. It ordered a new law and an early election, which was held in 2012.
For New Zealand: this is the closest precedent for 2026. New Zealand has no court that could do the same, so a fix has to come from Parliament, and the public has to ask for it.
Saxony-Anhalt 2026: MMP working as designed
Saxony-Anhalt has 41 electorates in a parliament of 83, so about half the seats come from lists. In September 2026 one party won 38 of the 41 electorates on 43.8 percent of the party vote. Under first past the post that would be an overwhelming majority. Under MMP it received 39 of 83 seats, three short of a majority, with no overhang.
For New Zealand: with half the seats on lists, a sweep of the electorates is absorbed. The Royal Commission proposed that balance for us in 1986, 60 electorates and 60 list seats. We now have 71 and 49.
Bavaria: one vote does not end it
In Bavaria both votes count towards a party’s total, so splitting your votes gains a party nothing. Even so, the largest party wins so many electorates that the parliament had 203 members in 2023, against a planned 180.
For New Zealand: an overhang does not need vote-splitting. Taking away the second vote would not remove it.
The same votes under three rules
How much difference does the overhang rule make? One way to see it is to take a single election and count it under different rules. The Gallagher index measures the P in MMP: the gap between votes and seats. Zero is a perfect match, and lower is better.
Baden-Württemberg 2026
| Rule | Size | Gallagher index |
|---|---|---|
| The state’s rule: add seats until balanced | 157 members | 0.36 |
| New Zealand’s rule today: keep the overhang, balance nothing | 134 members | 5.57 |
| Our fix: 120 votes, shared by party vote | 120 votes | 0.22 |
Schleswig-Holstein 2009
| Rule | Size | Gallagher index |
|---|---|---|
| The rule used: three overhang seats left uncorrected | 95 members | 1.98 |
| Seats added until balanced | 101 members | 0.86 |
| New Zealand’s rule today: keep the overhang, balance nothing | 80 members | 7.21 |
| Our fix: 69 votes, shared by party vote | 69 votes | 0.64 |
Our calculation from the official results. The index is counted among the parties that won seats, so that votes for parties below the threshold do not hide the effect of the overhang rule. Under our fix every electorate winner still sits, so there would be 134 and 80 members; the index measures votes in the House.
- Our fix comes closer to the party vote, in both elections, than either kind of balancing the German states use. And it does so without adding a single balancing seat: the votes in Parliament stay at the planned number, where Baden-Württemberg needed 157 members for a parliament planned at 120.
- New Zealand’s present rule is the outlier. On the same votes it is many times further from the party vote than any of the others.
- In Schleswig-Holstein the rule decided the government. Under the rule used, the governing parties had 48 of 95 seats. With seats added until balanced they would have had 50 of 101, and under our fix 34 of 69 votes. Neither is a majority.
New Zealand’s own 2026 result shows the same pattern. Counting all parties, the index is about 3.6 with the overhang and would be about 0.6 without it.
What the states do about an overhang
Every state with electorates deals with an overhang by adding seats for the other parties. Eight add as many as it takes. Five stop at a limit, which leaves the largest party with a small advantage.
- None turns electorate winners away. No state has copied the federal rule of 2023. All of them have preferred a larger parliament to an electorate without its MP.
- The price is size. Baden-Württemberg’s parliament is 31 percent over its planned size, and Hesse’s was 25 percent over in 2018.
- Close counts become contests. In Hesse and Berlin a single seat changed the size of parliament, and in Hesse a one-seat majority depended on it.
- None has tried our proposal. No state has MPs who sit without a vote, so the states offer no precedent for it, for or against.
What it shows for New Zealand
In parliaments of our size, an overhang is common and adding seats is costly. On 2026 election-night figures the same answer here would mean at least 146 MPs, with the Māori electorates set apart. And where an overhang did decide a government, it took a court to put it right.
Our proposal reaches the result the states want by another route. Every electorate keeps the MP it chose, as in every German state. Parliament stays proportional, as in every German state. And Parliament keeps 120 votes, which adding seats cannot promise.