Chapter Nine:
Citizen Initiated Referendums:
Adjunct or Antithesis of Constitutional Government?

Harry Evans

"It is better that a good measure should fail than that a bad one should be allowed to pass." 1


Citizen initiated referendums (CIR) is a generic term for related schemes whereby questions of public policy, particularly proposed laws, are to be submitted to the vote of all the electors in referendums, and thereby determined, at the request of a specified number of electors. These schemes are distinguished from the submission of questions to referendums by Parliaments or governments, in that referendums would occur automatically at the request of the specified number of electors, without governments or parliaments having any choice in the matter, and the questions in issue would be finally determined by the electors' votes, without any further action by Parliaments or governments to put the decisions into effect.

These schemes are regarded by their supporters as self-evidently good because they provide a more democratic method of decision-making, but they are also regarded as a renovation of the system of government.2

The purpose of this paper is to examine the relationship between CIR and constitutional government, that is, government conducted in accordance with constitutional rules and subject to constitutional safeguards against the misuse of government power.

For the purpose of this analysis, it is necessary to draw distinctions which were once clear but which have become obscured, and to refer to a debate which was decisively concluded in the past two centuries. The distinction is between, and the debate was about, democracy and representative government.

Democracy and Representative Government

In that past debate terms were used which then had reasonably clear and precise meanings, but which have since been largely emptied of meaning by misuse. One such term is "democracy", which has lost all established meaning through its appropriation by every type of regime known in modern times. The term once referred to a specific and distinct type of government, a system in which public policy issues, and particularly proposed legislation, are determined by the entire body of the citizens voting on such issues; in effect, government by referendums. This system was contrasted with representative government, whereby decisions are made by the chosen representatives of the citizens elected to institutions adapted for that purpose. There was then no misunderstanding that representative government is the same thing as democracy.

This distinction being clearly understood, there was in the 18th and 19th Centuries a debate about the relative merits of the two systems, a debate which was resolved decisively in favour of representative government. In order to analyse properly the relationship between CIR and constitutional government, it is necessary to revisit that debate.

Representative government was preferred in theory and flourished in practice because it was thought to have three great advantages.

In the first place, it was thought that democracy could be practised only with a relatively small number of people in a relatively small territory. Only representative government made possible popular participation in decision-making over wide territories and among large populations.3

This consideration may be regarded as being removed by developments in communications technology. As that technology develops, it will become possible for referendums to be held more often and on a larger number of questions. The spread of CIR itself following its introduction in Switzerland in the last century is largely a product of communications technology developments which have already occurred, beginning with mass circulation newspapers and progressing to television and networked computers.

Be that as it may, the other advantages which were urged for representative government as against democracy may still be regarded as worthy of consideration.

The second perceived advantage of representative government was that representation provided a quality control "filter" through which decisions were passed, and which allowed informed deliberation to be brought to bear on those decisions. It was thought that the people would elect as their representatives persons of above-average abilities and moral standards. Assembled in the legislature, these representatives would be able to deliberate, that is, to hear and be influenced by different arguments and to have their views formed and informed by that process. Such deliberation is difficult, if not impossible, with large numbers of people. Representative assemblies would thereby crystallise the informed view of the people rather than an uninformed view lacking the benefit of deliberation. The deliberations of the representatives would also help to form and inform the views of the citizens. Those views would ultimately prevail, but only when enhanced by the deliberations of their representatives.

"[It is the function of representative government] to refine and enlarge the public views, by passing them through the medium of a chosen body of citizens, whose wisdom may best discern the true interest of their country, and whose patriotism and love of justice will be least likely to sacrifice it to temporary or partial considerations. Under such a regulation, it may well happen that the public voice, pronounced by the representatives of the people, will be more consonant to the public good than if pronounced by the people themselves ...

"As the cool and deliberate sense of the community ought, in all governments, and actually will, in all free governments, ultimately prevail over the views of its rulers, so there are particular moments in public affairs when the people, stimulated by some irregular passion, or some illicit advantage, or misled by the artful misrepresentations of interested men, may call for measures which they themselves will afterwards be the most ready to lament and condemn. In these critical moments, how salutary will be the interference of some temperate and respectable body of citizens in order to check the misguided career, and to suspend the blow meditated by the people against themselves, until reason, justice, and truth can regain their authority over the public mind?"4

The third and most important consideration was that representative government makes constitutional government possible, that is, it provides the opportunity to subject government to constitutional safeguards.

The essence of constitutional government is that every authority should be forced to consider "whether its acts will be concurred in by another constituted authority", and should not be "exposed to the corrupting influence of undivided power, even for the space of a single year".5

If decisions are made directly by the whole people, how are any safeguards to be imposed on those decisions? What body can intervene between the decision-makers, the whole people, and the people who are affected by the decisions?6 If, however, the people delegate their powers to representative bodies, they can delegate different powers to different bodies and allow each to be a check on, and a balance of, the other. The allocation of different powers to different institutions is the primary form of constitutional safeguard, and where governments are chosen by the electors can be rationalised only on the theory that the electors, as a safeguard, do not entrust all their powers to any one institution.7

The primary form of the division of power in modern times has been the allocation of the executive, legislative and judicial powers to different bodies. Older than this device, however, is the division of the legislative power between two assemblies.

One of the prevailing themes of political thought throughout the ages has been that the legislative power, the power to make laws, the most fundamental and dominant power in the state, is too awesome and dangerous to be exercised by one body of persons alone, and should be divided between a number of duly constituted authorities, so that the consent of more than one body is necessary before the law can be changed.

The division of the legislature, by the institution of bicameralism, is almost a first deduction from the premise of representation :

"The history of mankind clearly shews, that it is dangerous to entrust the supreme power in the hands of one man. The same source of knowledge proves that it is not only inconvenient, but dangerous to liberty, for the people of a large community to attempt to exercise in person the supreme authority. Hence arises the necessity that the people should act by their representatives; but this method, so necessary for the support of civil liberty, is an improvement of modern times. Liberty however is not so well secured as it ought to be, when the supreme power is lodged in one body of representatives. There ought to be two branches of the legislature that the one may be a check upon the other. It is difficult for the people at large to know when the supreme power is verging towards abuse, and to apply the proper remedy. But if the government be properly balanced, it will possess a renovating principle, by which it will be able to right itself."8

Bicameralism, the requirement for proposed laws to be passed by two differently constituted assemblies, is not possible with democracy, that is, law-making by the whole people.

The other form of division of power which provides a safeguard is federalism. Apart from allowing the union of self-governing states without abolishing their separate governments, federalism has always been seen as a congenial method of dividing power consistent with popular control of the separate governments.9 Federalism and bicameralism as safeguards come together in the institutional device, adopted by the Australian Constitution-makers, of having the States in a federation represented in proportion to population in one house of the legislature and equally in the other.

All long-lasting systems of representative government employ these constitutional safeguards, the separation of the three categories of power, bicameralism and federalism, to a greater or less extent.

The question which arises is the effect of CIR on these constitutional safeguards.


CIR and Constitutional Safeguards


It is necessary to distinguish between two forms of CIR. The first, called the legislative petition referendum, or people's veto, involves submitting to a referendum, at the demand of a certain percentage of the electorate, a law already passed by the legislature, so that the electorate has an opportunity to veto that law.10 This amounts to adding another procedure to the process for making legislation, so that any law has another potential filter to pass before it can be enacted. This device may thus be regarded as an adjunct to constitutional government, because it adds an additional check or safeguard on the exercise of the legislative power. In effect it involves giving the electorate a veto over the works of politicians. Like all vetos, it may be used to frustrate good laws as well as foil bad ones, but that drawback applies equally to every division of legislative power, and every constitutional safeguard ever devised.

The other type of CIR, however, has exactly the opposite effect. Called the legislative initiative, it provides that any proposed law may be put to a referendum, and thereby enacted, at the request of a prescribed percentage of the electorate. This device of legislation by referendum by-passes the constitutional safeguards on the legislative power, and potentially places that power in one place, the electorate at large, defeating any division of the legislative power.

Either of these types of CIR can be used for changes to the Constitution as well as the enactment of ordinary legislation. In effect, Australia already has an enhanced version of the first kind of CIR for constitutional changes, because all proposed amendments of the Constitution must be put to a referendum. It is of great significance that built into this requirement for endorsement of constitutional change by referendum is a constitutional safeguard of a federalist character: proposed amendments, to be carried, must achieve the special majority specified by section 128 of the Constitution, majorities in a majority of States as well as an overall majority. This precedent for constitutional safeguards in referendums is referred to again below.

The second type of CIR applied to constitutional amendment would involve a proposed amendment being carried by a majority of the electors (whether a simple majority or a special majority) without the consent of the representative legislature.

It may therefore be concluded that the first type of CIR supports constitutional government by providing an additional safeguard on the law-making power, while the second type undermines constitutional government by bypassing such safeguards. In particular, the second type defeats the division of the law-making power.

It may be argued that bad laws may be passed however the legislature is constituted, and that no Constitution is proof against unwise legislation. The point is, however, that if such a law is passed by one house of a legislature, it may be rejected by another house of a different composition, particularly if there has been time for some sober consideration between the two deliberations and the two votes, or it may be significantly amended to cure its defects in the course of its passage through the legislature. A law enacted by referendum passes through only one process, one stage and one vote, with no opportunity for amendment, and with a terrible finality about it.

This is the basic problem with direct democracy: there can be no appeal from the whole people. It is this unique finality of the plebiscite which has made it so useful to dictators since the example was so well established by the two Napoleons, whose steps to absolute power were endorsed by referendums.


A large part of the case for CIR, however, rests on the failure of representative institutions and the consequent failure of primary constitutional safeguards.


Failure of Representative Institutions


It is now very difficult to maintain the two classical arguments for representative government which have been expounded above.

In relation to a representative assembly providing a filter, interposing deliberation between the people and decision-making, recent developments may be regarded as having destroyed this advantage. The rise of the professional politician, displacing the true representative who seeks the votes of the electors only after some other career, has notoriously resulted in legislators of somewhat lower calibre and standards than the average of the general population. To a person whose sole and life-long career is politics, winning elections rather than carrying out policies becomes the dominating goal. Partly because of professional politicians, legislatures have become less and less representative. The narrow bases and factional character of political parties are notorious and obvious in Australia. Intense party discipline, also most obvious in Australia, severely reduces the capacity of the representative to reflect the views of his or her constituents. As for deliberation, parliamentary debate has notoriously been supplanted by partisan shouting matches, blackguarding of opponents and scaremongering.

In relation to constitutional safeguards, party discipline has largely destroyed the legislature as such a safeguard. The theory of the British system of cabinet or responsible government was that the lower house would control the ministry; now the ministry, indeed, in Australia in recent times, the Prime Minister, controls the lower house. Bicameralism provides a partial safeguard only so long as the electorate returns different party majorities in upper houses, and the operation of such houses depends on parties which vote as blocs.

In these circumstances it is not surprising that CIR, even of the second type, is regarded as a corrective to degenerated parliamentary government. Part of the case for CIR is that the conventional legislative mechanism has become clogged by "debris", particularly in the shape of disciplined and factionalised political parties, and is therefore not able to function as it should. It is also said that conventional political procedures are too easily captured by vested interests and pressure groups, which may be defeated by referendums.11 Even the most uninformed votes by the people may produce better decisions than those of their representatives.

It is difficult to refute this case for CIR. Some counter-arguments may be raised. It may be argued that institutions which are not working should be reformed rather than bypassed, particularly as, in their reformed state, they would provide constitutional safeguards which CIR of the second type lacks. The substitution of a single-stage, simple process of legislating for a multi-stage, complex process may make it more likely that misconceived proposals, including those propounded by special interests, will slip through. Moreover, while it may justly be observed that parliamentary reform has proved slow and difficult, particularly as it involves breaking the stranglehold of the ministry over Parliament, the promotion of CIR as a remedy may have the effect of further retarding parliamentary reform. Such considerations, however, do not get around the fact that representative institutions do not achieve their stated advantages.


Effect on Judicial Review


One aspect of the effect of CIR on constitutional government which should be considered, and which has not been considered to any extent in the literature, is its effect on judicial review.

The phenomenon of increasing "judicial activism" is a feature of all systems of representative government, and has been particularly conspicuous in Australia in recent times. The High Court, not waiting for the conclusion of public debate over a bill of rights, has found implied rights in the Australian Constitution, and has expressed a willingness to change long established law when policy considerations appear to the judges so to require.12 The courts generally are more ready to review and overturn decisions of the political branches of government. This expanding judicial review is explicitly stated by judges to be a response to the failure of legislatures as safeguards of the rights of citizens.13

Where ordinary laws are enacted by referendum, it is still possible for such laws to be held unconstitutional by the judiciary. This has occurred in a few cases in the United States. For example, a referendum-approved State law to guarantee the right of persons to sell real estate to whomsoever they chose, i.e., to override fair housing laws aimed at racial integration, was overturned because it violated the equal protection provision of the US Constitution.14 In such cases the electors of one State are restrained from infringing minority rights contrary to the federal Constitution. The absence of CIR at the federal level avoids conflicts between the judiciary and the whole people.

In Australia the Constitution imposes fewer explicit limitations on law-making. The stated case for the discovery of implied rights is the protection of citizens against governments. It may be more difficult for the courts here to presume to protect the citizens against themselves, especially if CIR operates at the national level. Australian courts may then be much more reluctant to overturn referendum-enacted laws, as distinct from parliamentary enactments, on the basis of their inconsistency with the Constitution. Such action by the courts, particularly in relation to a law overwhelmingly carried at referendum, would be likely to generate considerable hostility to judicial review. Judicial restraint in respect of referendum-made laws would also protect parliamentary enactments, as the courts could hardly apply different standards of interpretation to the two categories of laws. In relation to amendments of the Constitution carried by CIR, judicial review of course provides no remedy, unless the judges interpret amendments contrary to their apparent intention, a course also perilous on which to embark.

CIR in Australia could, therefore, in the long run put a stop to judicial activism. On one view, this would be a good thing, as it would prevent unelected and unrepresentative judges usurping the legislative power. It could be regarded as restoring the constitutional balance upset by "judicial imperialism".15 It would probably not affect normal judicial interpretation and development of laws, as distinct from the discovery of hitherto hidden implications.

The adoption of a bill of rights, in conjunction with CIR, would give judicial activism a new charter. It could also lead to a head-on clash between the electors and the judges.


Building Deliberation and Safeguards into CIR


This analysis leaves a problem with the second type of CIR in relation to its effect on constitutional government, and the problem is still that set out in the classic case for representative government: the lack of deliberation and of constitutional safeguards in CIR.

The case against CIR as a means of enacting laws is basically the same today, although it is not often clearly articulated because it would be seen, correctly, as a criticism of democracy. The case is essentially that it would provide an opportunity for arbitrary and oppressive laws to be directly and finally enacted by an electorate seized by an "irregular passion", laws such as capital punishment for drug dealers, which would not be carried by a representative assembly, much less by two representative assemblies sharing the legislative power, and which, like prohibition in the United States, would be repealed only after great national travail and great damage to society. Referendums would provide a means for the speedy adoption of drastic and deceptively simple solutions to problems which lend themselves to appeals to such solutions. The situation is exacerbated by the replacement of newspapers as the primary source of information by television, with its "five second grabs" and its reliance on visual images.16

One of the few explicit statements of this case was made by a former Australian Senator at a seminar on the ailments of the current political system:

"I say that the most insane, unhelpful and destructive change that could be made to the nature of the Australian political culture is to introduce citizen initiated referenda (CIR). In my judgment, that would lead to the community permanently debating those issues which are the most socially divisive and difficult. There will always be CIR on the death penalty, abortion law reform and on those issues which are the most socially destructive and divisive, which should not, in my view, be worked out through that process. The real political issues, not the personal ones, will always be determined in a CIR framework on the basis that the largest quotient of ignorance will prevail."17

This may be regarded as a case against democracy, but it is also a case for representative government and the deliberation and safeguards which it should provide.

The question therefore arises whether it is possible to build provisions for deliberation and constitutional safeguards into CIR of the second type.

The introduction of CIR in Australia would require a constitutional amendment, whether CIR is to be applied only to ordinary legislation or to legislation and constitutional amendment.

The empowering amendment of the Constitution could entrench provisions of the following kind:

_ A requirement that proposed laws be published in draft form with a minimum period for public comment, so that any defects in drafting could be detected and amendments could be suggested.


A case can be made out for that special majority to be applied to ordinary legislation as well as constitutional change to be enacted by CIR at the federal level. The rationale for requiring that special majority even for ordinary legislation is that a similar special majority is already required for ordinary legislation through the equal representation of the States in the Senate. If it is thought that ordinary legislation should have a lower hurdle than constitutional amendment, some lesser requirement for geographical distribution could be required, such as at least 40 per cent support in four States, in addition to an overall majority.

Such provisions would go a long way towards overcoming the fundamental problem with the second type of CIR and providing for democracy with deliberation and safeguards.


An Adjunct


It may therefore be concluded that, subject to the kinds of provisions for the second type of CIR which have been outlined above, CIR in both its forms could be an adjunct to constitutional government, as well as providing for an alternative and democratic mode of decision-making.

There is still the problem of bypassing the representative institutions. It could be, as has been urged by supporters of CIR, that it would of itself lead to a kind of reform of the legislature, in that political parties in Parliament, with the constant threat of popular veto or popular initiation of legislative proposals, would conduct themselves in a more seemly and deliberate fashion.19

It must be remembered, however, that, for the foreseeable future, only a small minority of proposals would be enacted by CIR, and that representative institutions would still be relied upon for most law-making. The use of CIR will be limited by the amount of their time which people will allow to be taken up by public affairs. We should not, therefore, write off representative institutions and thereby forfeit their advantages altogether, but should persist with efforts to make them work as they should.



1. Cicero, De Legibus (On the Laws), III xviii 42, trans. C.W. Keyes, Loeb, 1977, pp.508-9.

2. The principal proponent of CIR in Australia is Professor Geoffrey de Q. Walker : Initiative and Referendum--The People's Law, CIS, 1987; The Rule of Law, MUP, 1988, esp. at p.390; The People's Law: Initiative and Referendum, University of Queensland Law Journal, 15:1, 1988; Let the People make the Laws, Australia and World Affairs, 9, Winter l99l; Current Australian Legislative Proposals for Direct Democracy, The Queensland Lawyer, 12:5/6, May 1992, p.190; Constitutional Change in the 1990's: Moves for Direct Democracy, Papers on Parliament No. 21, Department of the Senate, 1993; Address, Direct Democracy Seminar, Parliament House, Canberra, 1994 (transcript).

3. This is one of the theses of The Federalist No. 10, 1787 (James Madison), Everyman ed., p.45.

4. The Federalist Nos. 10 and 53 (James Madison), pp. 45-6, 322. The second passage refers to the Senate.

5. John Stuart Mill, Considerations on Representative Government, 1861, Everyman ed., pp.325-6. Significantly, these words refer to the necessity of bicameralism as a safeguard on the legislative power.

6. This point is made in The Federalist No. 10, p.45.

7. This theory was expounded by James Wilson: Lectures on Law, 1790-6, in The Works of James Wilson, ed. R.G. McCloskey, Harvard, 1967, pp.173-4.

8. Samuel Huntington, Speech to the Ratifying Convention of Connecticut, 9 January, 1788, in The Debate on the Constitution, ed. Professor B. Bailyn, New York, 1993, p.886.

9. The Federalist No. 51 (James Madison), p. 265; Lord Acton, May's Democracy in Europe, 1878, in The History of Freedom and Other Essays, MacMillan, 1922, p.98.

l0. The terms are Professor Walker's: Let the People make the Laws, op.cit., p.39.

11. Professor Walker: The People's Law: Initiative and Referendum, op.cit., pp.33, 45.

12. The judgments in the political broadcasts and Mabo cases, respectively: Australian Capital Television v. Commonwealth (No. 2) (1992) 104 ALR 389; Mabo v Queensland (No. 2) (1992) 175 CLR l.

13. Justice Gerard Brennan, Courts, Democracy and the Law, The Australian Law Journal, 65, January, 1991, pp.34-7; Justice John Toohey, A Government of Laws and not of Men?, Constitutional Change in the 1990s (conference), Darwin, 1992, p.15.

14. Reitman v. Mulkey (1967) 387 US 369. The US Constitution, art.4 s.4, contains a guarantee of "republican government" to the States. It has been argued that, as "republican government" encompasses representation, deliberation and safeguards, this provision could be used to overturn any referendum-made law on the basis that its passage violated these principles: Justice H.A. Linde, When is Initiative Lawmaking not `Republican Government'?, Hastings Constitutional Law Quarterly, 17:151, Fall 1989, p.159; When Initiative Lawmaking is not `Republican Government': the Campaign against Homosexuality, Oregon Law Review, 72, 1993, p.19.

15. This expression was first used by Professor Gordon Reid in 1979, long before the current outbreak of judicial activism: The Changing Political Framework, Quadrant, February, 1980, p.12.

16. According to Ms Pru Goward, ABC Radio National's principal political commentator, the "30 second grab" is obsolete and has been replaced by one of a few seconds: The Medium, not the Messenger, Senate Department Occasional Lecture, 20 October, 1995 (to be published in a forthcoming issue of Papers on Parliament).

17. Christopher Puplick, How Parliament works in Practice, Papers on Parliament No. 14, Department of the Senate, February, 1992, p.12.

18. A similar proposal to put deliberation back into American presidential elections is made by Professor J.H. Fishkin, Democracy and Deliberation, Yale, 1991.

l9. Professor Walker advances this argument: The Rule of Law, p.391.

Additional Notes:


The following additional notes on matters not mentioned in the paper may be of interest.

New Zealand: Indicative Referendums

New Zealand now has a system of indicative referendums, which are referred to as citizen initiated referendums, but which do not have one of the essential characteristics of CIR as set out in the paper.

The system, introduced by ordinary legislation in 1993, provides for referendums to be held on public policy questions at the request, made by way of petition, of 10 percent of registered voters. Thus it has the first characteristic of CIR, in that it is activated by electors without the approval of the legislature. The questions posed, however, are put in general terms rather than specific legislative proposals, and the votes in the referendums do not directly veto laws passed by the legislature or directly enact laws approved by the voters. The referendums are indicative only and not binding on the legislature, which still has the power to decide whether to implement referendum results by means of specific legislation. No doubt the government of the day, which effectively controls the single-chamber legislature, will be under strong electoral pressure to implement the results of referendums where the questions are passed.

An up-to-date account of the system is in "New Zealand's system of citizens initiated referenda", by Wayne Mapp, Agenda, 2:4, 1995, pp.445-454.

United States :

Constitutional Challenges to Referendum Decisions

Two decisions made by referendums at State level are currently the subject of challenges in the U.S. courts on constitutional grounds. One is a law passed by the voters of California to exclude illegal immigrants from access to welfare benefits. The other is a State constitutional amendment passed by the voters of Oregon to prohibit State laws preventing discrimination against homosexuals.

The grounds of challenge are that the laws violate the due process and the equal protection provisions of the federal Constitution. Both enactments have been held to be unconstitutional by lower courts. The two cases thus fall into the category referred to in the paper : State laws are challenged on the ground that they are contrary to individual rights guaranteed by the federal Constitution.