Referendums in the United Kingdom - Constitution Committee Contents

Memorandum by Caroline Morris, Senior Research Fellow, Centre of British Constitutional Law and History, Department of Law, King's College London

  1.1.  Member of the International Advisory Committee, IRI-Asia (Initiative and Referendum Institute).

  1.2  I make this submission in my personal capacity. My views should not be taken to have the endorsement of IRI-Asia or of King's College London.


  2.1  In a democracy, referendums are primarily employed to legitimate or substitute for government decisions.

  2.2  Referendums should therefore be binding on the government.

  2.3  If binding, the question to go forward must be carefully worded and of an appropriate nature for public decision of this sort.

  2.4  In particular, referendums should be held only on fundamental constitutional issues, with appropriate protections for fundamental human rights and those of minorities.


3.1  The role of the referendum

  3.1.1  The role of the referendum in a constitution is not always easy to pin down. Referendums may serve different purposes at different times. However, it is generally agreed in the academic literature that referendums are a "first-best" form of democracy for which representative democracy (the "second-best" form) attempts to substitute. This view is mostly premised on the importance of popular sovereignty and the consequent higher legitimacy that can be attached to decisions made by, or subsequent to, a referendum. Other goals served by referendums include the greater opportunities for participating in government by the people (and thus a possible decrease in political alienation and malaise) and providing a restraint on the powers of government.

  3.1.2  Despite these goals, referendums are not widely employed as a form of government. This is because, in my opinion, referendums are not well suited for determining complex questions of law and/or policy. This conclusion is borne out by the New Zealand experience of citizens' initiated referendums, to which I will return later. Referendums also carry with them the potential for manipulation of a government's popular standing, the abdication of government decision-making on controversial or difficult issues, and the danger that minority rights may be overridden by populist sentiment.

  3.1.3  They also suffer from logistical difficulties; the practicalities of holding a referendum, the cost, the appropriate timing (separately or with a local or parliamentary election) and the issue of how to respond to and/or implement the decision can steer governments away from their use.

3.2  Whether referendums should be indicative or binding

  3.2.1  New Zealand has legislated for indicative referendums, to be held on any topic (with a few restrictions) proposed by any individual and supported to go forward to a referendum by at least 10% of eligible electors.

  3.2.2  I do not support this approach. It has essentially led to the referendum becoming an expensive form of opinion poll, and the indicative nature of the result has led to frustrations with the government response. This increases the likelihood of political disengagement. The open-ended nature of the questions which may go forward has created a further set of interpretation difficulties, perpetuating the problem of arriving at a suitable response.

  3.2.3  However, where referendums are made binding on government, much more care needs to be taken with the wording of the question to be put, and the appropriate sort of question that may be put.

3.3  The appropriate subject for a referendum

  3.3.1  In a Westminster parliamentary democracy based on parliamentary sovereignty, a referendum may be held on any topic for which the empowering legislation provides. If the empowering legislation is general in form, there may be no restrictions, limited restrictions, or the holding of a referendum may be restricted to certain clearly delineated topics. I am of the opinion that this latter option is to be preferred.

  3.3.2  The New Zealand Citizens' Initiated Referenda Act 1993 chose the middle option: referendums may be held on any topic, other than:

    (i) calls for an inquiry into the way a previous CIR Act referendum was conducted or an electoral petition under the Electoral Act 1993; and

    (ii) those questions which have already been the subject of a referendum "of like effect" under the CIR Act.

  3.3.3  This open-ended approach has led to a wide variety of questions being proposed for referendum, of which four have gathered the required number of signatures for a referendum to be held under the CIR Act. The questions are:

    — Should the number of professional fire-fighters employed full-time in the New Zealand Fire Service be reduced below the number employed in 1 January 1995? (1995)

    Result: 12% Yes 88% No

    — Should the size of the House of Representatives be reduced from 120 members to 99 members? (1999)

    Result: 81.5% Yes 19.5% No

    — Should there be a reform of our Justice system placing greater emphasis on the needs of victims, providing restitution and compensation for them and imposing minimum sentences and hard labour for all serious violent offences? (1999)

    Result: 92% Yes 8% No

    — Should a smack as part of good parental correction be a criminal offence in New Zealand? (2009)

    Result: 12% Yes 87% No

  3.3.4  The first question came out of an industrial dispute, the second was a response to the recent increase in MPs following New Zealand's adoption of the Mixed Member Proportional Representation electoral system, the third followed a brutal attack on the proposer's elderly mother, and the most recent question followed the abolition of the defence of parental reasonable force in the Crimes Act.

  3.3.5  These sorts of questions (other than the question relating to the size of the legislature) are unsuitable as the subject of a referendum. Behind each question lies either an array of complex and involved policy decisions involving the careful weighing of different values and social priorities or a private matter rather than national public debate, which the public was not best placed to decide. These matters are not suitable for resolution with a simple "yes" or "no".

  3.3.6  To avoid this sort of scenario, consideration should be given to adopting one of two options for circumscribing the subject of a referendum. Obviously where legislation is drafted to provide for a single referendum question this would not be necessary, but if general legislation is adopted, some topic restrictions should be adopted.

The negative model

  3.3.7  Following the model of certain US jurisdictions, certain subjects may be made off-limits e.g. matters of taxation, defence, foreign affairs, the independence of the judiciary and constitutional rights. Likewise, the Venice Commission, in its Guidelines for Constitutional Referendums at National Level suggests that fundamental human rights, most notably "freedom of expression, freedom of assembly, and freedom of association" should not be vulnerable to the referendum process. There should also be a prohibition on questions that are of dubious value, such as vexatious, frivolous, scandalous, defamatory or indecent questions.

The positive model

  3.3.8  The alternative is to adopt a positive model, specifying in advance which subjects may be the subject of a referendum. In my opinion, the most appropriate topics are those which directly affect the constitutional make-up and powers of a state, ie the "constitutional issues" of the terms of reference. The 1975 EC membership referendum is an excellent example of the proper use of a referendum. Other appropriate topics would be a change to the electoral system or government (eg the adoption of a new voting system or the abolition of the House of Lords), devolved government, or changes to the sovereign powers of a state. Any alteration to the democratic fundamentals of a state should have the endorsement of its people. This aligns with the foremost rationale for the holding of referendums. This is something I believe was lost sight of when the NZ CIR Act was drafted.

  3.3.9  The protection of rights such as those contained in the Human Rights Act 1998 is somewhat more controversial but still possible.

3.4  The wording of the referendum question

  3.4.1  The New Zealand experience has shown this to be an interpretative minefield. Referendum questions should be clear, open to only one interpretation (e.g. what is a "smack"? what is "hard labour"?) and not be leading. To achieve this goal, questions should be carefully scrutinized before the enabling legislation is drafted. In New Zealand this function is performed by the Clerk of the House of Representatives. However, experience shows that the Clerk has been somewhat shy to engage with the proposed question wording. This may well stem from early litigation over the Clerk's actions over the re-wording of a question about battery hen egg production.

  3.4.2  There should also be a requirement that questions should address only one subject. The New Zealand legislation does contain a provision that questions should be capable of only one answer, but it does not seem to have been interpreted as allowing for the imposition of a "single subject" rule.

  3.4.3  For example, the 1999 criminal justice question can be broken down into five possible single questions. A "yes" vote for this sort of question is impossible to interpret: does the voter agree with the entire proposition, or only one aspect of it?

  3.4.4  I do not believe that voters are unable to follow and vote on a series of questions. For example, in the 1992 binding referendum on electoral reform, New Zealand voters were faced with a two-part question. In Part A they were asked whether they wished to change the electoral system. In Part B (regardless of their answer in Part A) voters were then asked to indicate their preference for one of four alternatives to the then First-Past-the-Post system. There does not appear to be any evidence that voters were confused by this format nor that it prevented them from expressing their views. In Part A, 1,271,284 people voted, and 92% of those voters then went on to vote in Part B (1 121 261 people).

3.5  Conclusion

  3.5.1  Referendums should be employed as a form of decision-making only where the usual channels of representative government are seen to be inappropriate.

  3.5.2  Because of this, referendums should be held infrequently, and when they are held, the result should be binding on the government. Frequent resort to referendums, or the use of indicative referendums, should be avoided.

  3.5.3  The binding nature of any referendum held necessitates careful attention being paid to the wording of the question which is to be put.

31 December 2009

previous page contents next page

House of Lords home page Parliament home page House of Commons home page search page enquiries index

© Parliamentary copyright 2010