Britain voted to leave the EU in a simple majority referendum. Gordon Bannerman (University of Guelph-Humber) argues that it would have done better to follow Canada’s example. After Quebec narrowly voted to avoid separation in 1995, the country revisited its approach to referendums. Indeed, not all UK referendums have involved a simple majority.
Canada has often featured on the LSE Brexit pages in terms of its past and prospective trading relationship with Britain. But surprisingly little comment has been made on the constitutional precedents Canada provided for pre-Brexit Britain, especially arising from the 1995 Quebec referendum, which resulted in a narrow 50.58 per cent to 49.42 per cent majority rejecting separation.
Afterwards, alert to the potentially momentous implications of a simple majority popular vote, government set to work to provide a new legal framework for referendums. Mistakes had been made. The federal government had no input in drafting the question, with the impetus handed to pro-independence parties (Parti Québécois, Action démocratique du Québec, and Bloc Québécois) whose question asked if voters agreed Quebec “should become sovereign after having made a formal offer to Canada for a new economic and political partnership”.
The convoluted, hedged language, described by the British High Commissioner to Canada Sir Anthony Goodenough as “gobbledygook”, concerning the future of Quebec/Canadian relations was not new. Even greater obfuscation had characterised the referendum of 1980, with a question consisting of 106 words, to grant Quebec’s government a mandate to negotiate “sovereignty-association” with Canada. In both cases, the language conveyed the impression of a smooth transition to independence amidst the willing cooperation and assistance of the federal government. The reality was that such a scenario could not be guaranteed. The Quebec nationalist parties had also threatened to make a unilateral declaration of independence (UDI) should a vote for separation not be honoured within one year. The Canadian government of Jean Chrétien, acting largely on the ideas of Stéphane Dion, the Minister for Federal Provincial Relations, requested the country’s Supreme Court to rule on the legality of UDI, relative to international law and Canada’s constitution. Acting on the 1998 judgment Reference Re Secession of Quebec, Chrétien codified the Supreme Court’s recommendations in the Clarity Act (Bill C-20) in 2000. Any future referendum question had to be presented in the House of Commons and approved by both Houses. Section 1(4) stipulated that questions providing only a mandate for future negotiations with Canada would be considered unclear and invalid. The question had to be “clear”, and voted by an (unquantified) “clear majority” for the Federal Parliament, as arbiter, to recognise its validity. No longer would 50 per cent + 1 be deemed sufficient for far-reaching constitutional changes – in future, a stronger mandate would be required. There the question has rested, as no referendum has taken place since 1995.
As some Canadian commentators have argued, Britain should have learned from the Quebec experience.
Some diplomats and politicians, such as Sir Anthony Goodenough and Malcolm Rifkind (Foreign Secretary, 1995-1997), did see important lessons from the Canadian experience but these were more pointed towards drawing parallels between the nationalist movements in Quebec and Scotland. While lauding the Canadian government’s measures in undermining separatism, the example of the Clarity Act in strengthening central government did not seem to gain traction in Britain.
Yet examples existed closer at home. Since the 1970s, referendums – previously almost unknown in Britain – have been used in popular decision-making. There has been no fixed formula. In the 1975 EEC referendum a simple majority was required, but the 1979 devolution referendums in Scotland and Wales, preceded by legislative acts, were framed more cautiously. Parliament was firmly in control of the process, and an amendment (Section 58(2) of the 1978 Scotland Act) by the Labour MP George Cunningham, supported by Conservative and Labour anti-devolutionists, stipulated that for the Acts to become law, an affirmative vote of 40% of the total registered electorate would be required as well as a simple majority.
The amendment imposed a qualified majority, by making the vote of every registered voter count – meaning non-voters were counted as votes against. The electoral register used was based on a qualifying date of October 1978. There were over 500,000 “unavoidable non-voters” including prisoners, hospital patients, and those moving home. Most notoriously, the amendment meant those registered in October 1978 who had since died were counted as votes against. The grievances were not all on one side—some argued the 40% rule reduced the anti-devolution vote, since many voters, by equating abstention with casting a vote against, stayed at home. The 40 per cent threshold was not attained in either country but in Scotland a majority of votes cast (51.6 per cent) voted for devolution but on a 63.72 percent turnout, the total amounted to only 32.9 per cent of registered voters. While making no difference in Wales, the amendment negated the simple majority achieved in Scotland.
The historical and political context was crucial in establishing the framework for the respective referendums. The 1974 Labour government, lukewarm on devolution but needing SNP parliamentary votes, reluctantly acted on the 1973 Kilbrandon Report’s recommendations, but readily accepted the Cunningham amendment as a safeguard. By contrast, the 1997 referendums for a Scottish Parliament and Welsh Assembly promoted by the pro-devolution Blair Government imposed no qualified majority rule. Despite the establishment of the Electoral Commission (EC) with referendum management, via the Political Parties, Elections, and Referendums Act (PPERA) of 2000 being statutory acknowledgment of referendums as a fixture in the UK’s political landscape, little has changed. In 2014 and 2016 the EC played a key role in ensuring intelligible wording of questions but it neither drafts original questions nor decides what constitutes a majority vote – those decisions are made by Parliament, and subject to the vicissitudes of party politics.
The simple majority formula used in the 2011 AV vote, Scottish independence vote of 2014 and the EU referendum of 2016 followed the format of 1975 and 1997 rather than the more cautious provisions of 1979. The assumption that the policy preferences of the main Westminster parties were shared by a majority of the electorate was reckless and dangerous. While validated in 2011 and 2014 (though for Scotland, as with Quebec, the result was perhaps closer than it should have been) the gamble did not pay off in 2016.
The Cameron government was remarkably complacent not only in setting the referendum terms but in failing to make any contingency plans (unlike in 1975) in the event of an EU exit – an extraordinary dereliction of duty, which, as the 2017 House of Commons report ‘Lessons Learned from the EU Referendum’ stated, led to unnecessary political instability: “Such preparation would negate the need for the Prime Minister to resign … It should be reasonable to presume that the sitting Prime Minister and his/her administration will continue in office and take responsibility for the referendum result in either eventuality”.
How far David Cameron was genuinely committed to the referendum promised in 2013 remains unknown. If a “bluff-call”, it clearly failed but in having to deliver a referendum it would have been wise to have considered the format more carefully. Qualified majorities are often used in the public and private sectors (not least in the EU), and in an electoral system underpinned by majoritarian principles, it is questionable how difficult it would have been to impose. While the proposal would probably have met with opposition from Eurosceptics within the Conservative Party, that doesn’t seem like a good reason for adopting a hazardous simple majority formula, lending credence to the view that Cameron was guilty of putting party before country.
Leading political scientists, including Vernon Bogdanor and Peter Hennessy, have expressed grave misgivings over using referendums, and the lesson for post-Brexit British governments seems to be to tread carefully. Before the 2016 referendum, Lord Kerr of Kinlochard was one of the few who found the absence of a threshold “constitutionally quite surprising for a decision as big as this”. Afterwards, of course, many Remainers concurred.
Britain’s non-codified constitution ensures referendums are considered legally advisory rather than politically binding but in practice, no government has been prepared to ignore the outcome of a referendum. Given these customary pressures, it makes sense to ensure a proper balance between expressions of the popular will while minimising the chances of a fierce populist outburst.
How different things looked in 1817, when George Canning stated:
“When I am told that the House of Commons is not sufficiently identified with the people, to catch their every nascent wish and to act upon their every transient impression,—that it is not the immediate, passive, unreasoning organ of popular volition,—I answer, thank God that it is not! I answer, that according to no principle of our constitution, was it ever meant to be so;—and that it never pretended to be so, nor ever can pretend to be so, without bringing ruin and misery upon the kingdom.”
By using referendums, we have moved far from Canning’s deliberative assembly model, but the Canadian example indicates that referendums can be – and should be – more closely managed. Canada’s Clarity Act might have been a useful reference point for British policy-makers, for protecting territorial integrity and promoting political stability are fundamental elements of responsible government. Imposing more rigorous conditions may seem Machiavellian to some, but it is far from incompatible with maintaining democratic legitimacy – in many ways it may ensure more mature and measured decision-making in the contemplation of far-reaching constitutional issues. The alternative can be decisions made on a transitory and/or ill-informed basis.
This post represents the views of the author and not those of the Brexit blog, nor LSE.
Gordon Bannerman is a professor at the University of Guelph-Humber.