A beginner’s guide to coalitions (NZ election)

The New Zealand election is still in a phoney campaign, with the parties only just beginning to launch their campaigns. Which gives me more time to fill in the background. Deborah has already done an excellent beginner’s guide to MMP, so I thought I’d follow up with what MMP naturally results in: coalitions.

Australians may think they are already familiar with coalitions. After all, you have one as a permanent feature of your political landscape, and it ran the country for eleven years under John Howard. But that’s a very traditional form of coalition. Here in New Zealand, we do things rather differently. The best way of illustrating this is to show how coalition arrangements have changed over the years.

Before MMP, there were no coalitions, or at least, none since the two-party system solidified in 1938. The unfair electoral system leveraged pluralities (and sometimes not even that) into majorities, allowing governments to do what they liked, with no checks and balances. Minor parties were an annoyance, not a potential partner in power, and so were generally ignored. As a result, New Zealand politicians were unused to coalitions when MMP came along. And the first MMP coalition, between National and New Zealand First after the 1996 election, reflected this. Management-wise, it followed the Australian model, with the two parties agreeing to act jointly on all things (I think they may even have shared a caucus, but I may be mistaken about that). New Zealand First gained the position of Deputy Prime Minister, as well as five Cabinet positions (and four Ministers outside Cabinet) – but those Ministers (and by extension, the party they dominated) were required to follow the Cabinet line in all things, thanks to the doctrine of Cabinet Collective Responsibility (this will become important later).

Thanks to past bad blood (Winston Peters, the NZ First leader, was a former National Party Cabinet Minister), the two parties did not trust each other, and so the coalition was governed by an exhaustive written agreement backed by a detailed agreement on policy. However, there were policy tensions between the parties, which were exacerbated when an internal coup within National saw it shift back towards neoliberalism. The coalition eventually fell apart, with National continuing to govern with the aid of defectors from NZ First and other parties – all of whom lost their seats in the 1999 election.

In retrospect, these sorts of arrangements were too brittle, and failed to deal with what is now called the “unity-distinctiveness dilemma” – the need for coalition partners to remain united for the purposes of advancing shared policies, but at the same time remain distinct so as to appeal to their separate bases of support. Every coalition arrangement since has been progressively looser in an effort to resolve this problem. So, in 1999, the new Labour-Alliance coalition (a minority coalition which needed support from the Greens to do anything) went for a much shorter arrangement, essentially agreeing on overarching goals and good faith discussion of how to achieve them. More importantly, it was recognised that the coalition parties would need to disagree on some things, and arrangements were made to allow this to happen without it causing the government to fall over. Most importantly, Cabinet Collective Responsibility was weakened (and the Cabinet Manual formally revised) to allow Ministers to disagree with the government in certain circumstances. In the end, this was not enough to save the Alliance – their poll numbers plummeted as they were seen to be “absorbed” by Labour, and the party eventually imploded due to tensions over the US invasion of Afghanistan.

The 2002 election saw even looser arrangements. Labour concluded a coalition agreement with the Progressives (a rump of the Alliance, centred on party leader Jim Anderton), but this was not enough to gain a majority on confidence and supply. So they cut a deal with the United Future Party promising specified policies and general consultation in exchange for support on confidence and supply. In addition, despite bad blood from the campaign, they also entered a cooperation agreement [PDF] with the Greens, promising general consultation and influence on policy development in certain areas. The post-2005 coalition agreements mirrored these sorts of arrangements, albeit with a different combination of parties, and with the innovation of support parties having Ministers outside Cabinet, bound by Cabinet Collective Responsibility only in their portfolio areas. So, we can have the Minister of Foreign Affairs criticise the government on free trade with China (not technically a foreign affairs topic) or immigration, without it impacting on their support for the government. The straitjacket has become very loose indeed.

While the experiment of Ministers outside Cabinet may not survive, loose coalition arrangements based on a few specified policies coupled with general consultation have been established as the norm. And the reason for this is simple: power. In a shared caucus, or where policy is decided on by a joint committee, small parties have no power. In the former, they can be outvoted. In the latter, at best, they can achieve deadlock. But by minimising their commitments, small parties maximise their freedom of action, and thereby their influence and ability to represent their constituents. Beyond a limited number of pre-agreed areas, the government has to come to them on every vote, allowing them to bargain for concessions, amendments, or (if they are inclined that way) quid pro quos. That said, the parties represent fairly well-defined policy positions, and can generally be relied upon to support legislation which advances their interests regardless. But the requirement for consultation gives them significant influence over policy before it even hits the House – and that is influence they will want to retain. I cannot imagine a return to a 1996-style coalition – small parties just won’t sign up to it. And if a large party insists on such an arrangement, they won’t get to be government.