One of the biggest bones of contention in the debate about MMP is ‘accountability’: most simply, the ability to ‘throw out the rascals’ when voters have had enough of them. Much of the debate is about accountability of governments, but accountability is also important for individual MPs. Following this line of argument, the Campaign for Change against MMP focuses on the ability of some defeated electorate MPs to return to Parliament on their party’s list. But are list MPs actually harder to ‘throw out’ generally, because of this ‘parachute’? And is it more difficult to ‘throw out’ MPs in general under MMP than under first past the post (FPP)? It may surprise many people, but the answer to both questions is ‘no’.

Many people who dislike MMP probably think that list MPs can stay in Parliament as long as they like, as long as they keep their ‘party bosses’ happy. But the job of being a list MP, particularly for a small party, is often not a secure one. If we define accountability in terms of MPs losing their seats, there has been far more accountability of this kind under MMP than there was under FPP. In 17 FPP elections since 1946, an average of 20% of MPs lost their seats or stood down. In four post-MMP elections (excluding 1996) the comparable rate was 25%. This is consistent with evidence from other countries: party lists generate greater voter-driven circulation in and out of parliaments when compared to what happens where candidates are elected from single-member FPP electorates.

Accountability of list MPs takes place through their parties. If voters do not like the performance of a party’s list MPs, they can withdraw their support. For this reason parties drop list MPs whose actions bring themselves and their party into public disrepute, as recent cases can attest. Replacements of list MPs between elections add a further element of fluidity. This may be as a result of scandal or allegations of illegal conduct, or simply because of retirement from politics. These shifts are additional to the voter or election-driven calculations of parliamentary change provided above.

Defeated electorate MPs who return on the list do attract a certain amount of popular ire. But these sentiments focus on a small number of cases. There are not enough of them to offset the greater voter-driven movement into and out of Parliament that takes place under MMP.

Research shows that most electorate MPs lose their seats because of swings against their party, rather than for their performance as MPs. Analysis of election results shows that some lose even though having increased their share of the vote compared to their result at the previous election. Most of these defeated MPs get more electorate votes than their party’s vote in the electorate, and some poll well above it. Over half of the defeated electorate MPs who have stayed in Parliament on the list did not contest the next election. Some have stood down before it. Some have refused to take a list position at the next election and have simply tried to recapture their electorate. Parties often rank defeated electorate MPs lower on the list at the next election, in less winnable positions. Some defeated MPs who retained a list seat have demonstrated continued local support by winning when they have contested subsequent local government elections. Some defeated electorate MPs who survive on the list do stay on longer: but they are a minority, and are valued by their parties and by many of those who vote for them.

If accountability is about the ability of voters to throw out MPs, MMP is no worse than the alternatives, and better than most of them. The MPs who are hardest to ‘throw out’ are those sitting for electorate seats that are traditionally safe for their party. The proportion of such seats would almost certainly grow under the first past the post, supplementary member, or preferential voting systems.

Further information documenting this analysis can be found in a paper downloadable here.

Jack Vowles is a Professor of Political Science at University of Exeter/University of Auckland.