The Check We Never Built

Part One: How Australia’s Information Environment Weakens Accountability

In Christ at the Centre, we explored how Madison gave us three checks on power, and how the Constitution quietly assumed a fourth: a people capable of seeing what their government does, and acting on what they see. We also introduced terra regius and terra populii, the Crown’s land that Australia legally remains, and the people’s land it was never built to become. This is the story of the fifth check nobody built to bridge that gap, and what happens to a free society when even the possibility of building it is quietly removed.

Where This Begins

My writing tends to focus on Australia. But this piece begins somewhere else, with a passage I have already quoted once in this series, and will return to again, because it will not let go of me.

In Federalist No. 51, one of a series of essays written in 1788 to persuade Americans to ratify their newly proposed Constitution, Madison wrote that if men were angels, no government would be necessary, and that the great difficulty of government is to first enable it to control the governed, and then oblige it to control itself. I wrote at length about that passage in Christ at the Centre, about Madison’s Augustinian honesty about human nature, and the system of checks and balances the framers built because of it. Executive against legislative. Legislative against judicial. Ambition counteracting ambition.

But Madison’s system assumed something else too, something so foundational that it barely needed stating, and so it wasn’t stated, not as a formal check alongside the others. It assumed that beneath all three branches, there was a fourth thing: a people.

A people who could see what their government was doing, form a judgment about it, and act on that judgment, at the ballot box, in the public square, in the slow accumulation of consent or its withdrawal that ultimately legitimises or delegitimises everything the other three branches do.

This is terra populii, the term I borrowed from the Reynolds/Tripp conversation and explored in Christ at the Centre. The people’s land. Sovereignty resting, in the end, not with the Crown, not with the parliament, not with the courts, but with the governed themselves, under God, as Locke would have it, and through him, the American framers.

Here is the question this piece exists to ask. What happens to that fourth thing, that unwritten, unappointed, structurally essential fourth check, when the mechanism by which the people perceive their government’s actions is itself degraded? Not seized. Not abolished. Just, degraded, quietly, by forces that are neither the government nor the people, until the judgment terra populii depends on can no longer be reliably formed?

That is the fifth check. The one nobody built into the system, because in 1788 the threat did not yet exist in a form anyone could have named.

A Contemporary Illustration

This is not abstract. Australia lived through a real-time test of the fifth check only a few years ago, and the result is already part of the public record.

During 2020, Victoria, like several other states, was governed for extended stretches by executive decision rather than by parliament. Victoria’s Legislative Assembly sat for 38 days that year, down from a pre-pandemic average of around 49. The annual figure, though, understates what happened when it mattered most.

A peer-reviewed analysis of how Australian parliaments performed during the pandemic found that Victoria’s Assembly met on a single day across the whole of April and May 2020, precisely the period when the first wave, and the government’s response to it, was at its most consequential. Curfews were imposed not through legislation debated and passed in Parliament, but through directions issued under the Public Health and Wellbeing Act 2008 (Vic), a power the Chief Health Officer can delegate to an ‘authorised officer’, a public servant, once a state of emergency is declared.

The curfew that affected millions of Victorians was, legally speaking, a decision made by one such official. When journalists and citizens pressed for the specific legal authority behind it, the answers were vague, then shifting, then, eventually, simply not given.

The Hotel Quarantine program, which detained returned travellers, including elderly, vulnerable, and mentally ill Australians, in city hotels, was later examined by a judicial Board of Inquiry, which found the program’s design, staffing and oversight fell well short of what the situation required. By the time that inquiry reported in December 2020, the second wave of infections it had helped cause was already over.

I covered this in full in Christ at the Centre, including the international parallel, Canada’s first-ever use of its Emergencies Act, in 2022, against the trucker protests, and the deeper argument about why a system with no settled answer to what a person is, and where their dignity comes from, finds it surprisingly easy to suspend that dignity once an emergency arrives. That argument stands, and I won’t repeat it here. As with everything in this piece, none of it required anyone to be a villain, only a system under genuine pressure, with nothing built in that required it to keep explaining itself to the people it governed.

What matters for this piece is narrower, and specific to the fifth check. During the months when Australians most needed to know what their governments were doing, and on what authority, the institutions whose job is to make that visible, parliament in regular session, courts and inquiries in real time, were either sidelined or arrived years late.

Terra populii depends on the public being able to see clearly enough, while it still matters, to form a judgment and act on it. For months at a stretch in 2020, that simply wasn’t available, and despite its implications for democratic accountability, relatively little public concern emerged about the reduction in scrutiny.

A Second Illustration — Who Gets to Decide What’s True

In September 2024, the Albanese Government introduced the Communications Legislation Amendment (Combatting Misinformation and Disinformation) Bill, legislation that would have expanded the powers of the Australian Communications and Media Authority, the regulator responsible for broadcasting and online content, to require digital platforms to police content the government considered “misinformation” or “disinformation,” with financial penalties for non-compliance.

The bill passed the House of Representatives in November 2024. It was then withdrawn after the Senate Environment and Communications Legislation Committee recommended it be discharged, with the Communications Minister acknowledging there was “no pathway to legislate this proposal through the Senate.”

I want to resist two temptations here, in opposite directions. The first is to treat the bill’s failure as proof the system worked and there is nothing further to discuss. The second is to treat the bill’s introduction as proof of a plan. Neither is right.

What the episode actually shows is something more useful than either: that a government, a regulator, and, implicitly, the platforms themselves were, for a period, converging on the idea that some institution needed the power to determine what counts as true enough to be said, and that this idea got as far as passing one house of Parliament before public and crossbench pressure stopped it. The Australian Christian Lobby was among the groups that raised concerns the framework could have been used against speech on contested questions of sex and gender, precisely the territory this series has documented elsewhere as already subject to coordinated institutional pressure.

The bill failed. But the conditions that produced it, a media environment increasingly concentrated, a government with substantial communications infrastructure of its own, and a genuine, real problem of online misinformation that someone, reasonably, wants to address, have not gone away. The next attempt may look different. The question terra populii poses is not whether this particular bill passed, but whether the public’s capacity to notice, understand, and respond to such attempts is itself intact.

It is worth being precise about how the mechanism would have worked, because the precision matters for this piece’s argument specifically. The bill would not have had government agencies telling citizens what they could and couldn’t say. It would have had a regulator telling platforms what they were required to police, with financial penalties if they got it wrong. The practical effect, for an ordinary person, may be that a post quietly stops being shown, or an account is quietly restricted, and what they experience is a platform’s decision, with no government fingerprint visible anywhere in it.

Madison’s checks worked because ambition could counteract ambition, but only because each branch could see what the others were doing.

Outsource enforcement to a private platform’s algorithm under regulatory duress, and that visible friction disappears: there is no branch left for a citizen to push back against, because there is no longer a decision anyone can point to.

That invisibility is itself a fifth-check problem: power exercised through unseeable channels is power the public cannot reliably perceive, judge, or contest while it still matters, which is close to the opposite of the transparency terra populii requires.

A Third Illustration — Closer to Home

There is a third example, closer to home, and it does not require us to take a position on the merits of any particular policy to see the structural point.

In March 2023, the South Australian Parliament passed the First Nations Voice Act, amending Section 3 of the state’s Constitution to establish a new representative body. It did this by ordinary parliamentary vote. No referendum was held, and for an amendment of this kind, none was required — South Australia’s Constitution can generally be amended by Parliament alone. This occurred a matter of months after Australians had voted on a comparable proposal at the federal level, with South Australians themselves recording one of the strongest “No” votes of any state.

None of this was unlawful, or even unusual within Australia’s constitutional tradition. State constitutions have long been amendable through ordinary parliamentary processes, and South Australia’s Parliament did exactly what its constitution allows, nothing about the process was irregular. The question this piece is asking is not about legality. It is whether arrangements of this kind, however longstanding and however lawful, sit comfortably beside the idea that sovereignty ultimately rests with the people.

I want to be careful here, because the point is not the merits of the policy either way. The point is the mechanism. A foundational document, the kind of document whose whole purpose is to define how power is exercised and limited, was altered without the people whose sovereignty it supposedly enshrines ever being asked. When elections for the new body were held the following year, turnout was recorded at approximately 9 percent, with some candidates elected on single-digit numbers of votes.

I am not asking whether the body itself is a good idea. I am asking a narrower and, I think, more important question: if a state’s constitution can be rewritten by the same people whose power it is meant to constrain, with no mechanism for the public to weigh in either before or after, in what sense does that constitution still constrain anything? And if a representative structure can be created and then draw single-digit participation, what does “representative” mean in that context?

The same point would hold for any future constitutional change in South Australia, in any direction, regardless of its substantive merits, which is exactly why the mechanism matters more than the example. A government that enjoyed this example could just as easily dislike the next one, and would find itself with no more recourse than the public has now.

This is terra regius again, in miniature, terra regius being the term this series uses for the legal reality that, in Australia, ultimate authority over the land and its laws rests not with the people but with the Crown, and in practice with Parliament acting in the Crown’s place. The Crown’s law, now Parliament’s law, went where Parliament willed it.

To be precise: South Australia’s constitution does reserve a small number of especially fundamental matters, chiefly, the existence of its own two houses of Parliament, for mandatory referendum. But recognising a new advisory body wasn’t one of those reserved matters, so no referendum was triggered, and under the ordinary amendment process, none was required. For this kind of change, which is to say, for most constitutional change, the people’s land remained, constitutionally speaking, notional.

The Documented Pattern

None of what follows requires a hidden hand. It requires only that we look at what is publicly reported, and ask what it adds up to.

Start with ownership. A 2024 international study tracking media concentration across fifteen countries found Australia has the second-highest level of media concentration in the world, behind only Brazil. In broadcast television, the top four players, Nine, Seven, ABC and Network 10, control 87 percent of the market, up from 79 percent. In radio, the top four, Southern Cross Media, ABC, ARN and Nova, control 77 percent, up from 57 percent. In newspapers, four companies control 84 percent of the market between them.

None of this is best explained as a plan to narrow the field. It reflects a genuine, global shift: advertising revenue that once funded newsrooms has moved largely to Google and Meta, while the cost of producing original accountability journalism, court reporting, freedom-of-information requests, long investigations, hasn’t fallen to match.

Outlets that can’t cover those costs merge, shrink, or close. It’s tempting to think the ABC and a more fragmented internet make up the difference, but the figures above already include the ABC as one of the ‘top four’ in both television and radio, and most Australians still form their picture of national politics through a small number of legacy outlets and the handful of platforms whose algorithms decide what those outlets’ content reaches them. A more fragmented internet doesn’t help terra populii if the fragments are still filtered through the same narrow set of gates.

This is not a conspiracy theory. It is a competition regulator’s nightmare, openly reported, and it has been getting steadily worse through ordinary commercial consolidation, mergers, acquisitions, the slow attrition of smaller and regional outlets that simply cannot compete. Twenty-five million Australians get most of their news, in one form or another, through a small number of corporate decision points.

Now add government. In the 2024–25 financial year, the Commonwealth spent $204.1 million on advertising campaign media placement, plus a further $61.2 million on campaign development — $265.3 million in total, delivered through a government-administered panel of approved communications suppliers. This is not improper. Governments have always needed to communicate with citizens. But it means that, alongside the commercial pressures already narrowing the field, the same handful of media organisations and the same panel of agencies are also among the largest recipients of government communications spending, a relationship that does not need to be corrupt to shape incentives.

Then add the think tanks. The pattern here is well established and not limited to any one side of politics. The Australian Strategic Policy Institute (ASPI), to take one well-documented example, receives $4 million a year in core funding from the Department of Defence, plus additional commissioned-report revenue from government that is not always itemised in annual reporting. ASPI’s research is then widely cited by the same media organisations as independent expert commentary, without always making clear that the institute producing it is substantially funded by the government department most closely connected to the policy areas its research addresses.

There is a fourth piece, and it runs in the opposite direction from the others, which is precisely what makes it worth including. In 2021, the federal government legislated the News Media Bargaining Code, requiring Google and Meta to negotiate payments to Australian news organisations for the use of their content. By the following year, those negotiations had produced an estimated $200 million in payments to local publishers, money that exists, by the government’s own account, because government intervened to create the conditions for it. Whatever its merits as policy, and there is a reasonable case for it, the effect is that a meaningful share of the revenue now keeping Australian newsrooms afloat depends on an arrangement government itself designed.

Again, no memo is required. No single actor needs to have planned this. What you have, when you put ownership concentration, government advertising spend, government-funded think-tank commentary, and a government-brokered revenue arrangement with the platforms in the same frame, is a closed loop: government communicates through a small number of outlets it also helps fund, directly, and now via the platforms too, those outlets cite research from institutes the government also funds, and the public receives all of this as independent journalism, independent expertise, and independent government accountability, three things that are, in this picture, considerably less independent of each other than many readers might initially assume.

A Road Not Taken

It is worth pausing on one historical “what if,” because it shows this gap was not inevitable.

When Australia’s Constitution was being drafted in the 1890s, one of its chief authors, Andrew Inglis Clark, borrowed an idea from the American Constitution: a guarantee that everyone would receive “equal protection of the laws” and “due process” before government could take action against them. In the United States, this kind of guarantee later became the foundation for courts to strike down laws that treated people unfairly or that gave government too much unchecked power over individuals.

Clark’s clause made it into the first draft. But over the following years, it was quietly stripped back, in large part because several colonies wanted to keep their existing laws restricting Chinese immigration and other racially discriminatory measures, and a broad “equal protection” guarantee would have put those laws at risk. By the time the Constitution was finalised, what remained was a much narrower, largely symbolic provision.

The point here is not to relitigate that history, which belongs to a different and difficult chapter of Australia’s story. The point is simpler: Australia came close to having a constitutional guarantee that would let courts act as a genuine backstop against government overreach, a check that works regardless of whether the public is paying attention or the media is reporting clearly. That guarantee did not survive. And its absence is part of why, when governments expand their own powers, during a pandemic, for instance, there is less for courts, and ultimately citizens, to stand on.

It is also worth being precise about what kind of failure this was, because ‘quietly stripped back’ makes it sound like an accident. It wasn’t only that. At the 1898 Convention, Clark’s clause was debated directly and voted down. Former South Australian Premier Alexander Cockburn argued that including such a guarantee would itself be an insult to the new nation, protesting that it would imply Australia’s own governments might otherwise ‘deprive any person of life, liberty, or property without due process of law’ (Official Record of the Debates of the Australasian Federal Convention, Melbourne, 8 February 1898, 688).

In other words, several framers didn’t just fail to build a judicial backstop — when one was offered, they argued Australia was too trustworthy to need it. The events of 2020, described earlier in this piece, invite a reconsideration of that confidence.

There is a postscript to this story, decades later. In 1992, without any change to the constitutional text, the High Court found that Australia’s system of representative government implies some freedom to discuss government and political matters, what lawyers call the ‘implied freedom of political communication’.

In one sense, it is the closest thing Australia has to the backstop Clark’s clause might have provided: a court-enforceable limit on what governments can do to the public conversation. But there are two limits worth noticing. First, an implied freedom constrains only government, it does nothing to stop a private platform from making the same decision, which is exactly the gap ‘A Second Illustration’ described. Second, when one of the legal challenges to Victoria’s 2020 curfew invoked this implied freedom directly, the Supreme Court found the curfew lawful regardless.

Even where Australia does have a version of the right Clark proposed, in 2020 it turned out not to reach the decision that mattered.

Structural, Not Conspiratorial

I used this phrase in Smart Cities, Smart Questions, and I want to apply it here with the same care, because it is doing real work and I don’t want it to become a slogan that loses its meaning through repetition.

These tendencies are not best explained by conspiracy or takeover. They are better understood as structural responses to modern pressures, complexity, liability, media cycles, the cost of producing journalism, and the fear of visible failure.

A media company facing declining advertising revenue merges with a competitor because that is what struggling businesses do. A government department facing a 24-hour news cycle and a hostile press gallery builds out its communications capability because that is what every modern government does.

A think tank facing the same funding pressures as everyone else takes government money because that is how think tanks are funded almost everywhere.

A regulator faced with a genuine flood of online misinformation reaches for the tools that regulators reach for, more power, more oversight, more capacity to define and act on the problem.

None of these decisions, taken on its own, is sinister. Most of them, taken on their own, are barely even interesting. The interesting thing, the thing that matters, is what happens when you add them together and let them run for two decades. Each actor, behaving rationally and often defensibly within its own incentives, contributes to an outcome that none of them necessarily intended and that may not serve any of them in the long run either: an information environment in which the feedback loop required for genuine accountability, citizens perceiving what government does, forming a judgment, and acting on it, becomes steadily harder to complete.

None of this is unique to government, or to Australia. The same convergence of concentration, complexity, and funding entanglement describes how citizens relate to large corporations, NGOs, and international institutions too, modern life is full of systems too large and too intricate for any one person to fully audit.

But government occupies a special place in this picture, for a reason this whole series keeps returning to: government is the one kind of institution whose legitimacy is supposed to derive, in the end, from the people it governs. A corporation doesn’t claim to rule by popular sovereignty. A government does. Which is why, when the conditions for popular judgment about government specifically start to erode, something more than inconvenience is at stake, it is the thing terra populii was supposed to guarantee.

To be fair, concentration on its own doesn’t guarantee uniformity. Journalists at the outlets named above continue to break stories that embarrass governments of every political colour. Independent media, podcasts, newsletters, and social media have genuinely expanded the range of voices available to anyone who goes looking for them.

The honest question isn’t whether dissent is possible, it clearly is. It is whether the average citizen, with a finite amount of time and attention, can realistically navigate an increasingly complex information environment without defaulting to the small number of institutional gatekeepers this piece has described. For most people, most of the time, the answer is no, not because they are incurious, but because nobody has the hours in the day to be their own newsroom.

I want to be just as clear about what I am not saying as what I am. I am not saying journalists are corrupt. The journalists’ own union, the Media, Entertainment & Arts Alliance (MEAA), is itself among the most vocal critics of ownership concentration, and for good reason; it affects their jobs and their craft before it affects anyone else. I am not saying ASPI’s analysts are paid liars, or that the public servants administering government advertising contracts are doing anything other than their jobs as defined. I am not saying any individual editor, minister, or regulator sat down and planned this convergence.

“Therefore judge nothing before the appointed time; wait until the Lord comes. He will bring to light what is hidden in darkness and will expose the motives of the heart.”  (1 Corinthians 4:5)

Scripture itself counsels exactly this kind of restraint. The task here is not to read hearts, that is beyond any journalist, and beyond me. The task is to bring what is observable into the light, and let the pattern speak for itself.

What I am saying is that the convergence is real, it is documented, and it has an effect, regardless of whether that effect was anyone’s goal.

Yes to a free press.

Yes to free markets, including in media ownership.

Yes to government communicating with its citizens.

Yes to philanthropic giving, think tanks, and global engagement.

And also yes, without contradiction, to transparency about who owns what, who funds what, and what the cumulative effect of all of it is on the public’s ability to see clearly.

These are not opposing positions. Pretending they are is itself part of how the gap stays open.

Where This Leaves Us

Before drawing this together, it is worth being precise about what is and isn’t being claimed. Elections still happen in Australia. Parliaments still sit, courts still operate, and the formal architecture Madison and the framers built is, on paper, intact.

The claim in this piece is not that any of that has been switched off. It is that the practical conditions required for those mechanisms to function as their designers intended, an electorate that can see clearly enough to judge what it is voting on, a parliament whose proceedings the public can actually follow, a citizenry capable of holding government to account between elections as well as during them, are not as robust as they once were. The forms remain. The question is whether the forms still do the work they were built to do.

Madison built three checks because he understood that power, left to itself, expands. He assumed a fourth, the people, without needing to build it, because in his context the people’s capacity to perceive and judge government action was not yet something that could be engineered away. It was simply there, in town meetings and pamphlets and the rough-and-tumble of an information environment too disorganised for anyone to fully capture.

That assumption no longer holds in the same way. Not because anyone seized it. Because the conditions that made it true have quietly changed, and almost nobody, not the media companies, not the government departments, not the think tanks, not most of the public, experiences this as a single event they could point to and resist. It happened the way Smart Cities argued societies usually lose freedoms: not through force, but through quiet normalisation, one rational decision at a time.

So here is where Part One leaves the question, and where Part Two will need to pick it up. If twenty-five million Australians form their political judgments inside an information environment shaped, not controlled, but shaped, by a handful of media owners, a government communications apparatus spending a quarter of a billion dollars a year, and a think-tank ecosystem substantially funded by the institutions it analyses, in what sense is sovereignty still meaningfully held by the people?

Terra populii was never fully built into Australia’s constitutional architecture, as Christ at the Centre explored. But even where it exists elsewhere, even where a constitution does formally rest power with the people, does it mean anything if the people cannot see clearly enough to use it?

A compromised information environment doesn’t just inconvenience the public. It acts like a cataract on the eye through which terra populii is meant to see, not blindness, but a clouding that makes everything technically visible and practically hard to make out.

“Then you will know the truth, and the truth will set you free.”  (John 8:32)

Scripture frames sight and freedom as inseparable. The promise here is not merely that truth is good to know in the abstract. It is that truth is the material freedom is made of. A people who cannot see clearly cannot, in any meaningful sense, be free, however many checks and balances their constitution formally provides.

How Australia’s Information Environment Weakens Accountability, was the subtitle. None of this points toward state control of the information environment, the abolition of media companies and think tanks, or the fantasy that citizens could somehow audit everything themselves. Part Two asks a more modest, and more open, question: whether mechanisms that place foundational questions back before the public directly, citizen-initiated referenda, which could extend a principle already present in section 128 of the Commonwealth Constitution, among them, might begin to do some of that work, even while the information environment around them remains imperfect.

That is the fifth check. Part Two asks what, if anything, could replace it.

"For nothing is hidden that will not be made manifest, nor is anything secret that will not be known and come to light." — Luke 8:17

The promise of a free society is not that power disappears. It is that power remains visible enough to be judged. Part Two asks whether Australia still possesses the tools required to bring that visibility into the light.