There appears to be significant interest in the issue of collective ministerial responsibility during Coalition government. For what follows, I will be largely quoting or paraphrasing Vernon Bogdanor’s The Coalition and the Constitution.
Following the formation of the Conservative-Liberal Democrat coalition in May 2010, the Coalition issued its Programme for Government which outlined in detail a full range of policy aims for the new government. It also provided for the explicit abandonment of the doctrine of collective responsibility on one issue in particular – the referendum on the Alternative Vote. The two parties would be whipped to get the Bill implementing the referendum through the House of Commons, but would be free to campaign on opposite sides during the referendum itself.
Elsewhere in the Programme for Government, there were explicit “agreements to differ” on issues such as new nuclear power stations, the renewal of Trident, on married couples’ tax allowances and on university tuition fees.
Agreements to differ were not inventions of the current Coalition. They were first introduced in 1932, to hold together the National Government on the issue of free trade. As Bogdanor writes:
Three Liberal Cabinet ministers had been allowed, not merely to abstain, but to speak and vote, both in the House of Commons and in the country, against the Import Duties bill, which imposed a tariff, being promoted by their Conservative coalition colleagues. The ‘agreement to differ’ was defended by Stanley Baldwin, Lord President of the Council and in effect deputy Prime Minister, as an expedient for a coalition since ‘[t]he fate of no party is at stake in making a fresh precedent for a National Government’. But, he went on to warn, ‘[h]ad the precedent been made for a party Government, it would have been quite new, and it would have been absolutely dangerous for that party’. (pp 51-52)
The doctrine of collective responsibility was suspended twice in the 1970s, not by a coalition government, but by a Labour government, both times on issues involving Europe. Consequently, ‘agreements to differ’ aren’t limited to coalition governments, but are more likely to be needed by coalition government than by single-party government.
What is different about the 2010 agreements to differ is that they are the first to be agreed to before the formation of a government, as part of a Coalition Agreement, rather than after the government is formed, as an expedient way to deal with new and/or unforeseen issues. Bogdanor states that because of this, “the ‘agreements to differ’ of 2010 are, from a constitutional point of view, more soundly based.” (p. 52)
What does an agreement to differ do? Simply put, it allows for the suspension of the convention of collective ministerial responsibility for certain specific issues. Interestingly, Bogdanor posits that there is no reason in principle why a Cabinet – any Cabinet – should not agree to suspend collective responsibility on any issue. The New Zealand Cabinet Manual has accepted the legitimacy of the agreement to differ since 2008, thus governments in New Zealand are free to waive it whenever they wish. Bogdanor does not note, however, that New Zealand has been governed by coalitions since the adoption of Mixed-Member Proportional in 1996, which may explain why agreements to differ are specifically recognized in the Cabinet Manual.
Bogdanor explains that with an agreement to differ, what is actually happening is that the doctrine of collective ministerial responsibility is replaced by collective Cabinet unanimity – in other words, the Cabinet collectively agreed to suspend collective responsibility. Ministers are therefore collectively responsible for the decision to suspend collective responsibility on a particular issue for a limited time.
Bodganor also makes clear that on none of the occasions when past governments have used agreements to differ did it bring credit to the government in question. All it succeeded in doing was to draw attention to important differences within the government. He concludes his discussion of agreements to differ thusly:
Not only a government but any collegiate executive, whether a Board of Directors or the committee of a students’ union, will make itself look ridiculous if it publicly advertises its disagreements. Any such body must therefore weigh up the benefits of employing an agreement to differ rather than trying to discover a compromise formula on which all can unite. It is surely advantageous if a coalition Cabinet can discover such a compromise formula. For the sanction on any departure from the convention of collective responsibility is not constitutional but political. It lies in the danger of public ridicule. (pp. 53-54)