MMP may deliver a parliament which reflects us, but frequently the government does not.

At the heart of my recent history of New Zealand, Not in Narrow Seas, is the interaction between economic and social change. I could measure economic change via the – far from comprehensive – economic data base. Social attitudes were much harder, for there are less data.

There are written records about what people are thinking, but coverage is confined to writers and is selective. This social statistician wanted to use social surveys but they hardly exist before the 1970s. A common strategy is for the historian to impose their views and while I could not avoid that for myself – no one can – I wanted to minimise my personal view as much as possible.

I thought I had found an ingenious solution. Every three years or so we have general election, which is a kind of survey whose single question ought to throw light on our political attitudes. I knew there would be limitations from tactical voting, but hopefully the results would indicate some trends.

Covering elections is standard historical practice, of course. It would be a rare general historian who did not mention which party won a particular election and draw, often explicitly, inferences about the tone of the nation. What I was going to do was the much rarer approach of looking at party votes won rather than party seats won. There is a sense that Not in Narrow Seas is the first MMP general history.

I added one further wrinkle. It is usual to report on each party's share of the vote. The history uses share of total enrolment. That is because non-vote fluctuates from election to election., Bob Chapman, who was a pioneering professor of Political Studies, argued that the `non-vote party' sometimes determined electoral outcomes; ignoring it can be misleading. In 1951 the Holland-National Party was returned with an enhanced majority, despite having a reduction in its share of enrolled voters; Labour's share fell more. The reality of the Waterfront Dispute was not that the population turned its back on the wharfies - many turned their back on parliament.

When I looked at the voting data, I was surprised how often the number of seats did not reflect the party vote and hence the composition of the government. Having lived through the 1978 and 1981 elections I knew that Rowling's Labour party won more of the popular vote than Muldoon's National party, which became government because it still won more seats. (As an aside, those who grumble that Donald Trump became president of the United States on a minority of the popular vote overlook that a similar thing happened here – twice in their lifetime.)

I knew too, of odd outcomes in other elections, like the fact that despite Coates' Reform Party winning more votes in 1928 than Ward's United Party, the latter formed the government. What surprised me was how often there were such eccentric outcomes. It is arguable that had the vote on the right not split among three parties in 1935, the Forbes Coalition Government would have retained power; that would have rewritten the course of our history.

The same sort of thing happened in 1993 when the left vote split between Labour and the Alliance, Had that not happened, the Bolger National government would had been a short-lived three year term and Mike Moore would have returned to the premiership.

That does not happen so much today under MMP (although there is still some tactical voting because of the threshold provisions). However, there is another peculiarity of our political system.

The choice of government in 2017 was between National-NZF and Labour-NZF-Green, other options being ruled out. Each was some distance from the political centre. The first coalition was about 0.8 standard deviations to the right of the New Zealand on a left-right scale and the second about the same distance to the left. That meant that about 85 percent of the population were to the left of the first coalition (including those at the centre) and to the right of the second. On the social conservative-progressive scale the first coalition was also about 0.8 SDs on the conservative side, while the second was a slightly closer at 0.65 SDs on the progressive side, so that only 75 percent were less progressive than the government ruling them.

The sharp change in the politics of the government did not reflect a sharp change in the politics of the voters. Certainly there was a move to the left between 2014 and 2017 (of about 6 percentage points) but it was not as dramatic as the change in the politics of the government.

The old electoral system was called FPP - first past the post - a silly name as there was no post; one could win an electorate with less than 40 percent of the vote. I once called the voting system 'Winner Takes All' but I decided while writing my history that 'Front Runner' (FR) was an even better description, since whoever was in front on election day took the seat, even if they had 40 percent or less of the vote.

MMP is not perfect, but it does produce a parliament which reflects the demography and ideology of voters reasonably closely, much more so than front-runner parliaments did. However, parliament selects the government on a winner-takes-all basis rather than a proportional one. So, unlike parliament, the government does not look like us.

(Moreover, parliament is not very good at holding the government it appoints to account - that argument belongs to a later column.)

Not in Narrow Seas recalls that, once upon a time, the elected government would decide in the interests of, say, 80 percent of the voters, because it feared it could not be sure that at the next election it could hold all the seats it won if it did not. You could argue that under MMP the government only needs to target a smaller proportion of voters. I am not saying that it always does. Sometimes we elect statesmen and stateswomen who do try to act in the nation's interest.