Explaining the demise of the national–New Zealand first coalition

Abstract

This article considers the fall of New Zealand's first coalition government under the Mixed Member Proportional (MMP) electoral system, through the prism of a ‘unified’ model of Cabinet durability. This model incorporates both the ‘structural attributes’ of the government at the time of its formation, as well as the stochastic ‘critical events’ that occur over its lifetime. However, the conventional attributes of the National–New Zealand First coalition were considered favourable, and the critical event that precipitated its demise did not by itself appear to be a sufficient condition for termination. It is argued that the hitherto unspecified structural attribute of intra‐party stability was a significant factor for this particular coalition, and that this case also demonstrates the cumulative impact of non‐terminal events in undermining the incentives for continued cooperation between coalition partners.

Notes

Peters probably sought to avoid being outvoted by the larger party, and by walking out of the meeting, he denied it a quorum according to the coalition agreement, which stated that ministers from both parties must be present. However, the agreement also adopted the Cabinet Manual as the basis for the functioning of the coalition Cabinet, which has a different interpretation of the quorum rule. For a more detailed analysis of this issue, see McLeay (1999).

Earlier that day another NZF MP, Deborah Morris, had also resigned from the party to sit as an Independent. She later resigned from Parliament, and under the Electoral Act, the next candidate on NZF's party list entered Parliament as a representative of that party despite Morris's defection, boosting its caucus to nine.

The coalition officially terminated on 26 August 1998, after a notice period of seven days.

Minimum winning theory derives from rational choice models of decision making which assume that actors select from among alternative strategies those which they expect to maximise their individual preference orderings, while at the same time minimising costs. A coalition is in equilibrium in the sense that no alternative governing arrangement is preferred at the point of formation.

Rather than assume that all politicians are motivated by self‐interest, the concept of utility upon which actors based their decisions was modelled instead as a collective benefit based on the agreement by the members of a party on a set of polices.

The importance of the number of formation attempts may be heightened in PR democracies where governing options are traditionally canvassed in turn. However, the rejection of Labour by NZF did not indicate great difficulty in forming a workable government, since NZF had the luxury of considering another formation attempt simultaneously. Brinkmanship was also a factor, as Peters went sailing with his caucus on the day after the election, and did not begin talks with National until five days later (and with Labour two days after that).

The approach to Labour was probably a ploy to guarantee the terms of the coalition agreement, particularly the mid‐term promotion of further NZF MPs to Cabinet, given the tough talk of the new National leader. Labour leader Helen Clark was invited to join with NZF under the terms of the agreement made with National, despite the fact that Labour had been unprepared to match this offer in 1996. The chances that Labour would agree when it was in a much more powerful position, and NZF in a much weaker state, were remote. Clark stated her preference for an early election, and in this eventuality, NZF would have faced electoral oblivion.

Support for NZF slumped from 13.4% on election day to 2% in opinion polls following the formation of the coalition, where it continued to languish throughout its time in government.

If the Prime Minister's majority is in doubt (a more likely scenario under MMP), the Governor‐General could conceivably refuse a request for dissolution and move to appoint an alternative government from the existing membership of the house. However, the use of such vice‐regal discretion runs the risk of charges of partisan bias and political interference against this non‐elected office.

This argument could be carried further to situations where parties are undisciplined, or at one extreme where they do not exist at all, leading to the expectation that Cabinets will almost always be short lived. New Zealand's political history before the advent of parties in 1891 would substantiate such a claim (see Church 1998).

For example, the divided fourth Labour government (1984–90) had three different leaders in its final 18 months, yet only one Labour MP broke away to create a new party. Splinter parties under FPP were limited by the knowledge that a new electoral player would find it difficult to win seats.

The loss of support was probably more serious for NZF—unlike National, it was at risk of not only losing office but also losing parliamentary representation altogether.

Coalition studies consider a government to have terminated when it resigns or is defeated in parliament, adds or loses a member‐party, changes its majority status, changes its leader, or incurs an election—regardless of whether it wins or loses (Warwick 1999, 378).

It could also be argued that the coalition became less connected as a result of Shipley's accession to the National Party leadership, as her own ideological positioning lay to the right of her predecessor.

A change in the decisive structure of government between elections may arise from such splits and fusions, constituting a de facto decision rule shock, yet this is not considered by Laver and Shepsle (1996, 210).

For example, ACT expected to be consulted on all matters once it confirmed its support for the minority government on confidence and supply.

Shipley was either lucky or had shrewdly calculated her actual support. Two NZF MPs resigned before the outcome of the Dispute Committee meeting, and a NZF caucus meeting immediately afterwards confirmed the departure of five more MPs. The final MP to defect did so on his return from an overseas trip.

Electoral volatility may have also encouraged NZF to believe that it could recover enough votes to survive.

Despite the fact that the Reform caucus had itself atrophied to a parliamentary minority by this stage.

Jonathan Boston is Professor of Public Policy in the School of Government at Victoria University of Wellington. He was a member of the New Zealand Political Change Project (1995–2003) which examined the behavioural, institutional and policy impact of the introduction of proportional representation in New Zealand in 1996. His main research interests lie in the areas of public‐sector management, comparative government and tertiary education funding. Stephen Church works in the New Zealand Parliament as an adviser to the United Future Party. After completing his PhD at the University of Canterbury he was appointed as a research fellow for the New Zealand Political Change Project based at Victoria University of Wellington, where he also taught political science and public policy, from 1999 to 2002. Hilary Pearse is a graduate student in the Department of Political Science at the University of British Columbia. Her research interests include political institutions and political culture in Western democracies, with a particular focus on parties and electoral systems.

Early elections are relatively rare in New Zealand's electoral history.

However, the disputes to which multiparty governments are susceptible may in fact be little different from those experienced by some of New Zealand's recent single‐party majority governments. The difference arises in the reduced ability to conceal dissension behind the cloaks of Cabinet and caucus secrecy, when the government is already divided across two parties. Given the long history of single‐party majority government, parties and voters alike may find it difficult to accept such differences as healthy signs of robust debate within the government. The National–NZ First coalition exhibited remarkable cohesion in parliament even though its internal difficulties were well known. Despite the clause in the Labour–Alliance coalition agreement that allowed the parties to ‘agree to disagree’, the coalition continued to uphold collective Cabinet responsibility until the 2002 election.

It would be tempting to draw a connection between this pattern of devaluation and rising hazard rates for terminations, except that the increase in risk over time applies more accurately to governments ending in early dissolutions rather than a replacement, as occurred in this case, where the risk is more constant.