The Price of Finding Out: The Third Condition of a Just Democracy

The third condition of a just democracy asks nothing about what citizens believe. It asks what it costs them to find out, and who sets that price.

A citizenry can be wrong and still be free. It stops being free when finding out becomes somebody else’s decision.

The Just Democracy framework holds that a government becomes legitimate only by meeting conditions prior to any vote. The third is an informed citizenry whose political choices are not systematically manipulated by fear, money, or ideology. A citizen must be able to find out what those who govern would prefer they not know, by a route controlled by nobody with a stake in whether they find out. A citizen must also have reason to know there is something to find out, because a question nobody thinks to ask is as closed as one nobody may ask.

Fear, money, and ideology are the vectors that raise the price. They are not the test. A prison term is fear, concealed funding is money, a flooded record is ideology. Manipulation here is influence that works around a person’s judgment rather than through it, and systematic manipulation is that influence built into what reaches people rather than attempted one campaign at a time.

A route can be closed at the far end as well as the near one. Where someone with an interest in the answer works to make the record untrustworthy, they charge the price after the reading rather than at the door. Russia is furthest down that road. The method RAND analysts named a firehose of falsehood in 2016 does not argue but exhausts, drowning the record in repetitive and mutually contradictory claims across every channel rather than suppressing it. What separates that distrust from earned distrust is the arrangement that produced it.

The test is what it costs a citizen to find out. Every state charges something, so the size of the price does not settle anything, and the question is who sets it. Friction that is nobody’s decision, falling on everyone because societies are large, is the ordinary cost. Friction somebody chose to impose on the record as a class is the failure, whoever’s interest it serves. Transparency, as Onora O’Neill argued, discharges nothing if those it is made for cannot assess it. Amartya Sen put it concretely: famines do not occur in functioning democracies, where a free press and an opposition can raise the alarm. Singapore tests this hardest. It is among the least corrupt and best administered states in the world, and it has a ministerial power to order corrections published, newspaper companies whose controlling shares cannot move without the government’s consent, and no law granting access to official records. It fails the condition, and failing it while governing well is the point rather than the embarrassment.

Walter Lippmann stated the problem in 1922: citizens act on the pictures in their heads of an environment “too big, too complex, and too fleeting for direct acquaintance.” Whoever shapes the pictures shapes the politics. A citizenry that defers to them has not consented. It has complied. The Islamic tradition met the problem as one of evidence rather than politics. Ibn Khaldun opens the Muqaddimah asking why so much of what passes for history is false. Reports get believed because they flatter a party, or because nobody examines who carried them. The hadith scholars had built the institutional answer, not a body ruling on which reports were true but a published method, usable by anyone trained, for tracing who had carried a report.

Pakistan raises the price with fear, and has written down who sets it. In January 2025 Parliament amended the Prevention of Electronic Crimes Act to make it a crime, punishable by three years in prison, to spread information the state judges false and likely to cause fear, panic, or disorder. The Act created an authority empowered to order material removed within 24 hours, and widened the class of persons who may be defamed to include institutions of the state. That last provision is the confession. A rule protecting an institution from false statements gives the institution a complaint against its critics, and the state the power to rule on it. The Ministry of Interior constituted the authority in March 2026. Parliament set the price the day it wrote the office into law.

The United States raises the price with money, and it did so after promising it would not. In Citizens United in 2010, the Supreme Court struck down limits on independent political spending by five to four, and upheld the disclosure requirements by eight to one. The majority made the second holding the ground of the first. Transparency, Justice Kennedy wrote, “enables the electorate to make informed decisions.” The spending arrived. The visibility did not.

Social welfare nonprofits may spend on politics without naming their donors, and since a rule change in 2020 the IRS no longer collects those names either. The money reaches campaigns through committees that do disclose, and disclose that a nonprofit or a shell company gave it. By the Brennan Center’s count, published in 2025, undisclosed money in the 2024 cycle exceeded $1.9 billion, more than $1.3 billion of it routed that way. That made 2024 the most secretive election since the ruling that promised transparency. The Center says its own figure understates the total, because some channels cannot be tracked at all.

The remedy exists on paper and cannot fire. The Federal Election Commission needs four votes of six to act, and told Congress in 2019 that it had deadlocked in more than half the enforcement matters it had considered since 2012. It lost its quorum in May 2025. The DISCLOSE Act, which would require the disclosure the Court assumed, has been before Congress since 2010 but has never passed.

An American reader will by now assume they know which coalition this describes. Neither. OpenSecrets counted more than $514 million of undisclosed money supporting Democrats in the 2020 cycle against roughly $200 million supporting Republicans, and Democratic aligned groups drew more of it in three consecutive cycles. The party whose legislators have almost entirely sponsored the disclosure bill has gained most from its failure. The objection would be identical if the money ran the other way, and for three cycles it has.

Canada is the most uncomfortable case. A law written to prevent the damage did it instead, and did it with none of the three vectors. Bill C-18, the Online News Act, took effect in December 2023, on the premise that platforms carrying the reporting for nothing should pay for it. Google bought a five year exemption from the regulator in October 2024 by funding a journalists’ collective. Meta declined to bargain and blocked Canadian news on Facebook and Instagram instead. It was leaving news everywhere, and Facebook’s referrals to a tracked sample of news sites worldwide fell by about half in the year to March 2024. What the law changed was the shape of the withdrawal in Canada: total rather than gradual, immediate rather than drifting, aimed at one country’s journalism. The block has held in substance ever since.

The Media Ecosystem Observatory (MEO) reported a year into the block that Canadian outlets had lost the great majority of their engagement there and roughly a third of local outlets had gone dormant on social media. What replaced the link was the screenshot. An earlier count from the same researchers found posts carrying screenshots drew the engagement news links once had. A report with its chain of transmission cut is a report nobody can check. Their own conclusion is that what fills the space is systematically different from the journalism it replaced. MEO’s June 2026 survey found that only 41% of Facebook users and 26% of Instagram users knew Canadian news organizations can no longer post there. They have not left the platforms. Some 53% of all Canadians use social media for news, more than the year before. No censor was appointed, no journalist arrested, every newspaper remains a search away, and most of the country’s social media users are scrolling past an absence they do not know exists.

That the law was well intended is beside the point. A government that degrades the information environment while trying to repair it has degraded it, and a citizenry cannot consent to what it cannot see. What keeps Canada a recoverable failure rather than a permanent one is that Parliament can repeal what Parliament passed, the regulator publishes its decisions, and the researchers who measured the damage may publish the measurement. That is capacity rather than performance. Australia passed a comparable law and shows what performance looks like. Facebook blocked news there in February 2021. Canberra amended the law within the week and the block ended. When the deals it produced lapsed, Canberra legislated again. Three years into the block, Canada has turned none of its capacity against the block: the Act stands unamended, and Ottawa has been negotiating with Meta rather than compelling it. A mechanism that has never been made to fire is a claim about a democracy, not evidence about one.

Sweden shows what the opposite looks like. Public access to official records has sat among the fundamental laws since 1766, not in a statute a government can amend when disclosure becomes awkward. Any person may demand a public document, needs no reason, need not be a citizen or a journalist, and ordinarily gets it promptly and for little or nothing. The record is open unless a specific rule closes it, and the burden of naming that rule falls on the official who refuses. The concession is also the proof. The exemptions for security and privacy are broad, and the parliamentary ombudsman keeps finding agencies too slow to comply, which is a record of citizens demanding documents and an institution ruling on the refusals. The mechanism fires. Sweden does little about the wider environment. It makes the record itself cheap, which is what this condition asks for.

One objection carries real weight. Manipulation does not work as advertised. Joshua Kalla and David Broockman, in the American Political Science Review in 2018, pooled forty field experiments and added nine of their own, concluding that the best estimate of the effect of campaign contact and advertising on American voters’ candidate choices in a general election is zero. Hungary is the same finding. In April 2026 voters removed a government that had spent sixteen years building the media environment they voted in. If persuasion is that weak, worrying about manipulated citizens looks like contempt for them. But those studies measure campaigns trying to change minds in the weeks before a vote, the smallest part of what this condition covers. The condition asks about the standing environment, not whether a message moved a voter, but whether the voter could have checked it, and at what price. That price is measurable downstream. In studies from 2020 and 2022, American municipalities that lost a newspaper paid five to eleven basis points more to borrow, and penalties for regulatory violations at local facilities rose by about 15%. Those figures measure what officials and firms do when nobody is watching rather than what citizens found out.

The other objection is that any standard of an informed citizenry hands somebody the power to define informed. The condition survives that by being procedural rather than substantive. It requires no citizen to hold any belief and takes no position on which beliefs are correct. Mill’s argument in On Liberty is usually cited for the speaker’s freedom and is stronger about the listener: an opinion never contested becomes dead dogma rather than living truth, and a people never required to defend what they believe cannot tell believing from knowing. So the condition asks about the route rather than the destination: whether the record exists, whether reaching it is anyone’s to permit or withhold, and whether anyone is working to discredit it in advance. A citizenry reaching the wrong conclusion with that route open has met the condition. One reaching the right conclusion with the route closed has not, and will not reach the next one.

The three conditions this series has examined fail together more often than alone. Law that bends for the powerful cannot compel disclosure from them, an armed establishment outside civil authority decides what the press may report about it, and a citizenry that cannot find out casts a ballot without knowing what it is voting about. Pakistan charges liberty and has named the office that sets the price. The United States hides the receipt, though its own highest court told it that the receipt was the price of everything else. Canada never sends the bill, which is why most Canadians have not found out their news is gone. Sweden settled the question in 1766 and has been charging next to nothing since. What separates these four is not what their citizens know, but who decided what knowing would cost.