The Voice Referendum: Cheating the Constitution

The first Australian referendum to amend the Constitution was held in 1906 and the short question to be addressed in the ballot box was:

Do you approve of the proposed law for the alteration of the Constitution entitled “Constitution Alteration (Senate Elections) 1906?”

If you didn’t know what this statute (“proposed law”) was going to do in respect of Senate elections, too bad—you didn’t matter. The same form was adopted for the next thirty or so constitutional referendums: it was a substantial disfranchisement of many Australian electors each time.

This essay appears in our Voice online-only edition.
Subscribers and non-subscribers alike can download
the entire August special issue by clicking this link

Section 128 of the Constitution provides the lawful way of amending the Constitution, but the understanding of this section is made difficult, even to lawyers, by its mixing its essential thrust with long auxiliary provisions (such as the relation between the two houses of Parliament). These obscure a very clear essence, but when in our reading we excise them, the following simple and now obvious central structure comes into focus:

The proposed law for the alteration thereof must be passed by an absolute majority of each House of the Parliament, and after its passage through both Houses the proposed law shall be submitted in each State and Territory to the electors.

When a proposed law is submitted to the electors the vote shall be taken in such manner as the Parliament prescribes…

…And if in a majority of the States a majority of the electors voting approve the proposed law, and if a majority of all the electors voting also approve the proposed law, it shall be presented to the Governor-General for the King’s assent.

The key to understanding the essence of our referendums is submission. It is perfectly clear from these provisions that a law not submitted to the electors cannot be approved by them and therefore cannot be presented for the King’s assent and cannot amend the Constitution.

The short question in the Voice referendum process—the one to be submitted to us in the ballot box—reads as follows: “A Proposed Law: to alter the Constitution to recognise the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice. Do you approve this proposed alteration?”

The parliamentary bill that establishes the referendum sets out this short question and then goes on to state a longer version of the “proposed law”; this is the actual constitutional text that would be inserted into the Constitution were the referendum carried. This longer statement adds two things of substance to the short question: first the Voice’s function (of making “representations to the Parliament and the Executive Government of the Commonwealth …”) and second, the Voice’s structure (Parliament’s “power to make laws with respect to matters relating to the Voice, including its composition, functions, powers and procedures”).

These two substantive provisions (making representation and structure) could certainly be inserted into the Constitution if they were “submitted in each State and Territory to the electors” and were carried in the terms of section 128. But the simple fact is that they are not being submitted by the present referendum and therefore cannot possibly be validly inserted into the Constitution. What is to be submitted to us is the short question; but it is not planned to insert that into the Constitution.

The government has tried to hide anything about what the Voice is to be doing—the question that has occupied the whole debate so far. Most people want Australia’s first peoples to be recognised in the Constitution; and the government is trying to get them to vote on that alone—a vote that would carry the hidden matters with it, a veritable constitutional fraud.

It is doing this in two ways: (a) removing what the Voice is to be doing and how it is to be structured from the submission to the electors in the ballot box, and (b) making an illicit delegation of those powers to Parliament, as explained in an earlier article (“Unconstitutionality in the Voice Amendment”, Quadrant, June 2023). It is attempting to create a new and unique constitutional entity that it only partly defines, leaving the rest (delegating it) to Parliament.

The power to construct the Voice is given to the legislative power of Parliament. Now, the conferral of a new legislative power (as was achieved in the 1946 Benefits referendum, and attempted in the 1951 Communist Party referendum) is not a delegation of the amendment power to the Parliament, it is the creation of a legislative power. The difference with the Voice referendum is that it is intending to create a new and unique constitutional entity that it only partly defines; it is leaving (delegating) the rest of the definition to Parliament—hiding it from the referendum.

A law not submitted to the electors cannot be approved by them and therefore cannot amend the Constitution. The formal submission by Parliament is in its short question of the proposed law to the electors. But could there not be another form of submission constituted by the whole activity of the referendum, including particularly the Yes and No cases mailed out to electors? The Parliament would in this way, by establishing the complex referendum process, be submitting it to the approval of the electors. There are several reasons that preclude this.

First, it would be difficult to tell what was a satisfactory extrinsic process and what was not (readers will remember that the Prime Minister was for a long time against allowing the Yes and No cases). Second, it would still be necessary to look also at the actual question in the ballot box which might by its precision be taken to limit an elector’s understanding of the extrinsic process—there is no way the formal question can be ignored here. Third, the process of the referendum is intrinsically formal, involving the creation of a proposed law starting in the form of a bill and ending if approved in an Act of Parliament. This fits the formality of the ballot box, but not the controversy and wheeling and dealing in the politics of the referendum.

Statutes (and their bills) have a long title and a short title, the latter usually, but not always, shorter than the long. I shall not distinguish between these two titles, and simply refer to the title chosen for the short question in the referendum law.

In the first referendum of 1906 the proposed law was referred to by its title (the proposed law for the alteration of the Constitution titled “Constitution Alteration (Senate Elections) 1906”. And this perfunctory practice of submitting simply the title of the law continued for more than fifty years. The practice makes no sense at all. By the time “a proposed law is submitted to the electors” (the wording of section 128) the Parliament has done its work and the issue turns to the sovereign people. There is no point in expressing the people’s issue in terms of a Parliamentary enactment that most people will not even know where to find. It’s not as though Parliament’s work in creating the referendum would drop away if not asserted—if the electors approve, it goes straight into the Constitution without question. For a valid referendum there must be a submission by Parliament and an act of approval (or not) by the electors; but the perfunctory form is neither a submission (information is held back, not submitted) nor a possible approval (approval requires knowledge of the thing approved). And so these early referendums were obviously invalid as inconsistent with section 128.

The title is a mere distraction—completely irrelevant to the constitutionality of the vote. The proposed law might be titled “Albo and the Vibe”, and it would be judged not as a title, but as a guide to approval or not of the long statement. (And would fail.) A real (indisputable) title—say, Bill number xy of 2023—raises a different issue. This is a proper name that refers precisely to what is to be inserted into the Constitution, but as an issue for approval or not, is completely meaningless to an elector who knows nothing of that bill.

In 1967 the titling of the proposed laws became longer, more descriptive and more informative:

Do you approve the proposed law for the alteration of the Constitution entitled
“An Act to alter the Constitution so as to omit certain words relating to the People of the Aboriginal Race in any State and so that Aboriginals are to be counted in reckoning the Population”?

This was a crucial change. But though more informative—as this one is—titling raises the possibility of being less informative in that it gives the Parliament the opportunity to do precisely what has been done in the Voice referendum: choose between elements of the proposed law to exclude those believed to favour a No vote.

There was now no necessity to worry about the question of title. So, in the four referendums of 1977, instead of referring to the title of the amending statute the amendment itself was described, for example (the Territorial Voting referendum):

It is proposed to alter the Constitution so as to allow electors in the territories, as well as electors in the states, to vote at referendums on laws proposed to alter the Constitution. Do you approve of the proposed law?

This was a perfectly fair submission by Parliament for the electorate’s approval (or not). But this new descriptive practice, too, contained the possibility of abuse, choosing between elements of the proposed law to exclude those that are thought to favour a No vote.

In 1984 the Referendum (Machinery Provisions) Act (48,793 words!) reinstated title, requiring the short question to “set out the title of the proposed law”. What the Parliament did to deal with this problem in the Voice referendum was to satisfy the requirement of title by quoting the title (short or long) that the bill itself is given. But who gives the bill its title? The Parliament does. And so this title, too, is open to exactly the same No vote abuse.

That there be, following a successful referendum, a precise statement of what is to be inserted into the Constitution is obviously necessary (the referendum itself cannot be the place for argument about detailed wording), and it must be the Parliament that makes this precise statement and publicises it. But to inform some electors of it—those who read the bill—is not to submit it even to them. The issue in the Voice referendum is not one of detailed wording: the function and constitution of the Voice is a major part of the very essence of the referendum. To submit these essential elements to the electors is not hard to do. The Republic referendum was much more complex than the current one, but its short statement (remove the Queen and install a Parliamentary-elected President) caught the substance with accuracy and so its being submitted to the electors could count as a submission of the longer form.

It is not the difficulty of stating the substance that has moved the present government: it is its wish to remove two issues of real contention from its submission to the people. It seems to have feared it might lose the referendum on these real and more complex substantive issues (issues that have dominated the political debate). Issues not submitted are not capable of approval in the terms of section 128—“submit” and “approve” (or not approve) are logically correlative.

In Boland v Hughes, Chief Justice Mason said, “it is for Parliament to decide what shall be the content of the proposed law to be submitted to the electors”. He was referring both to the short question and to the long version in the whole bill. Parliament has a significant discretion as to both formulations. But there is a big difference between them. The discretion as to the long version is unbounded, whereas the discretion as to the short question is a discretion within the necessary requirement that it count as a submission of the long version (whatever it is) to the electors.

Section 128 also requires the Parliament to choose the manner in which a vote shall be taken, but that too is on the assumption that the long version is being submitted to the electors.

Many previous referenda have been simple and short, where there was no reason for not allowing the whole long version to be the short question (here the perfunctory practice’s silly choice of playing with “title” has now caught up with it—how can the title be the same thing as the content?!)

The Republic referendum was aimed at a complete rewriting of much of the Constitution. For complex referendums there must be a question short of the whole proposed text. The short question (call it X), is the only formulation of the bill that is actually submitted to the electors, and so the question after it is: how does the long formulation of the proposed law (Y) get to be submitted to the electors so that it can be inserted into the Constitution? X is to be submitted to the people, but Y is to be put into the Constitution. There must be a valid connection. And the only possible connection is this: X must be a short and fair description of Y (as it was in the four 1977 referendums). Only then is it the case that by submitting X to the electors, Y also has also been validly submitted to them.

There is no doubt that here there is here a wide discretion in the Parliament as to this connection; but, still, it is the issue. Of the Voice referendum, we must be able to say that: “Recognising the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice” (X) is a short description of: “Recognising the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice that may make representations to the Parliament and the Executive Government of the Commonwealth on matters relating to Aboriginal and Torres Strait Islander peoples; and recognising also that Parliament has power to make laws with respect to matters relating to the Voice, including its composition, functions, powers and procedures” (Y).

No matter how wide the discretion of Parliament, it is not even faintly arguable that this X is a short description of this Y. It follows that the Voice referendum fails on the issue of its submission to the people and is invalid.

It’s not as though there is difficulty in drafting an X that is a short statement of Y: “A Proposed Law: to alter the Constitution to recognise the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice that is (a) constituted by Parliament and (b) may make representations to the Parliament and the Executive Government of the Commonwealth. Do you approve this proposed alteration?” is short and would do.

The Australian Constitution, which guarantees the basic citizenship of all races, genders and creeds, is the peak of Australia’s constitutional law. How is it that many Australian lawyers (and even Bar Councils) have not in the matter of the Voice referendum seen the Constitution as their first allegiance? This is by no means to say that they should not as individual citizens support the Voice, only that their professional integrity should attach to the Constitution, not to a particular sectional interest. But are they not supporting the Constitution when they advocate its amendment in the Voice referendum? This is a spurious claim. On the question of amendment their professional integrity must attach to the lawful way of amending the Constitution, not (except personally) to a particular sectional amendment. And here, the many lawyers advising the government have overlooked the failure of the referendum to submit the true questions to the electors because their attachment to the cause of the Voice has clouded their professional integrity. This, regrettably, is the case with the larger Bar Associations, whose principal function is the professional integrity of lawyers. Many lawyers do practise that professional integrity, but when senior barristers and judges call them racist for doing so, it is their own professional integrity that they trash.

And there are simple mistakes. Some lawyers have missed the invalidity of the current referendum because they have been misled by the absoluteness of the amending power. They (correctly) see that section 128 is unlimited in the kind of amendment that can be proposed by Parliament (the absolute power of the Parliament to define the long statement) and because of this absoluteness ask no questions. These lawyers overlook the fact that the absoluteness itself is conditional on Parliament’s submission to the people in the ballot box’s short question. They misunderstand the people’s absolute sovereignty to which the Parliament must submit.

The referendum process that Parliament has established for the Voice is a “matter arising under the Constitution and involving its interpretation”, as to which the High Court clearly has jurisdiction to strike down an unconstitutional referendum result after the event (as Mason CJ said in Boland v Hughes). The structure that Parliament has given to the Voice referendum is invalid under section 128 and an abuse of, not merely a mistake about, the Constitution.

Now, judgments of public administrative law sometimes come to the conclusion that government action whilst legally wrong is nevertheless valid. It would be a very big thing for the High Court to invalidate a successful referendum; and so might it not say that the short question put to the electors in the Voice referendum was legally wrong, but nevertheless the referendum result is valid? No, it might not.

As I have said, the process of the referendum is intrinsically formal, involving the creation of a proposed law starting in the form of a bill and ending if approved in an Act of Parliament. And Acts of Parliament are never legally right or wrong; they are either valid or invalid. Is the Act constituting the Voice referendum, and therefore the Voice referendum itself, valid or invalid? It is invalid. Should the High Court be approached on this, we could expect it so to rule.

Michael Detmold is Professor of Law emeritus at the University of Adelaide.