The rule, exactly. Before an election each citizen chooses: to vote — or to take a payment and not vote in that election. The payment is a percentage of the median income, set by referendum; there is no fixed sum. One ballot is one vote, with no multiplier in the count; the vote is amplified only in that each ballot's share grows as others step out. The budget pays under law; a candidate never pays. Only a referendum of all citizens — simple majority, no quorum — introduces, changes or repeals the rule.
The protocol has been introduced nowhere and no pilot has been run: shares, turnout and outcome figures in the chapters are estimates, and the protocol promises nobody an election victory. If a chapter says otherwise, Exact Answers and the Charter are correct. For a candidate: ten questions and ten steps. For a citizen, a mayor, a finance officer, a donor, a journalist, a scholar, a lawyer: answers by role. Everything in force in one file: llms-full.txt.
Levitsky and Ziblatt: Gatekeepers and the Drained River¶
Chapter: 06 — Structural Hypocrisy and the Arsenal of the Elites File: 06_039 · v1 · 12 August 2026 (session 37)
Their frame¶
"How Democracies Die" (2018) stands on two legs. The first: modern democracies die legally, at the hands of the elected — capture of the referees (courts), strangling of the media, rewriting of the rules; four behavioural markers of an authoritarian at the entrance (rejection of the rules of the game, delegitimation of opponents, toleration of violence, readiness to curtail freedoms). The second, for them the main one: the last line is not the constitution but two unwritten rules of the elites: mutual toleration (the rival is not an enemy) and institutional forbearance (do not squeeze everything out of legal powers). The formula of the death of Weimar and Venezuela: it is not that the populist is strong — the gatekeepers are weak: parties, in fear of defeat, themselves let the authoritarian into the mainstream.
Their three blows at AB-EXIT¶
1. Demobilisation of civil society. "You pay for exit from politics; when the authoritarian goes for the seizure, the streets will be empty — everyone took the cheque." Answer (034, 056c): button B demobilises the ballot box, not the street — the rights of protest are untouched; and an attempt at seizure under AB-EXIT switches on the one protest that everyone joins: a payment taken from 55–65 % of the population (winter fuel × the whole country). The mechanism does not disarm society — it arms it with a common, instantly verifiable trigger.
2. A dual-use tool. "The markers are clean (does not reject the rules — a referendum; does not delegitimise opponents; expands freedoms), but a payment register + the precedent 'money for electoral behaviour' is a gift to the next authoritarian." This is Fraenkel's syringe — closed by safeguards of constitutional level (034) and self-verification of a violation (038): selective manipulation of the payment is the only crime of the state provable by every victim in five seconds.
3. "You are treating the wrong address." "Democracy is surrendered by the gatekeeper-elites, not the voter. Hitler was brought in not by the electorate — he was brought in by von Papen, a deal of elites. How will your button B help against a pact of party bosses with the devil?" — The strongest blow; the answer below, through a recount of their own premise.
A recount of their theory¶
AB-EXIT strikes not at their conclusions but at the initial premise, which they accept without defence: the populist wave is an independent variable, an element; the people generates authoritarians, the only question is the strength of the door. Hence their hopeless prescription ("parties, be responsible") — a request to the dam to be stronger under an uncontrollable flood.
The recount: the wave is not an element; the wave is a product. It has a cost price (mobilisation of the aggrieved by cheap lies) and a profitability — both controllable by a mechanism. Button B buys out the authoritarian's mobilisation base before the gatekeepers' door: at 55–65 % exit a critical mass of the manipulable physically does not gather in the ballot box. In their terms: not strengthen the gatekeepers — drain the river before the dam. Their central drama ("in the moment of panic the elites surrender") loses its stage: panic happens when the authoritarian "cannot be ignored" at 25–35 % — and he no longer gets that; von Papen does not make a deal with a man who has 6 %.
The second recount: forbearance acquires a price. Their "unwritten rule not to squeeze the powers" rests on pure norm — a violation has no price. The thermostat (13.9, Axelrod 13b.3) builds one in: squeezed → the median sagged → the dividend froze for everyone → the return of the avengers with double weight. The unwritten rule for the first time gets a written enforcement mechanism.
The third: their own historical material works for the mechanism — each of their cases began with the authoritarian cheaply buying the aggrieved (Chávez — oil populism, Orbán — national grievance, Trump — the rage of the rust belt). AB-EXIT is a counter-bid at the same auction: the state outbids the price for the same voter, legally, by formula, forever.
Orbán as a control case (28 September 2026)¶
The thesis heard in every conversation about "the death of democracies": through ordinary elections, on the Orbán pattern, hidden autocrats arrive, with a sleeping majority on security populism they demolish the institutions and drift into autocracy — and the protocol protects against this "with a large margin". The core is right and is already in this file (the wave is a product; the base is bought out before the door). "With a margin" — no. A recount on Hungary.
Two strata of the sleeping majority. Fear is not a stake but an induced state; only the one who loses personally has a stake (001b §3). The populist deliberately breeds the suggestible and deliberately extends the vote to them, knowing they are easier to manage: in 2010 Orbán gave citizenship and the vote to ~1M ethnic Hungarians abroad — postal voting, 94 % for Fidesz, ~300,000 votes. For rural Transylvania the dividend is a month's wage, and there is no stake in Hungary at all. The reserve the autocrat built with his own hands leaves first.
The Fidesz-2022 breakdown (54 % of votes at 69.5 % turnout = 37.5 % of the electorate), by income and dependence on the budget, not by stated motive:
| stratum | % of electorate | remains |
|---|---|---|
| clientele: public works, KESMA, state contracts, party activists | 7–9 | 7–9 |
| concrete fear: the 2015 border counties, villages near reception centres | 2–3 | 2–3 |
| Hungarians abroad | 2–2.5 | 0–0.5 |
| abstract fear (Soros, Brussels) + need: villages, pensioners — those who were paid the 13th pension (029.4b) | 23–25 | 5–8 |
| total | 37.5 | 14–20 |
The opposition, 24 % of the electorate — urban, better off, votes by voice rather than by fear — keeps ~70 % → 17; Mi Hazánk 4 → 2. Result: Fidesz 42–51 % instead of 54 %. The two-thirds vanishes in every scenario, a simple majority is in doubt. Everything turns on the row "abstract fear + need" — whether 5 or 8 remain; the test: how many of them took the 13th pension and went to vote — that is the upper bound of those whom fear will hold without money.
The door the protocol does not close. In 2010 Fidesz took 52.7 % of votes at 64 % turnout — and 68 % of seats: the two-thirds came from the mixed formula, not from the voter, and the constitution was rewritten with that norm. The protocol counts votes, not mandates; at 45 % and the same formula Fidesz can still take more than half the seats. To Levitsky's third blow ("you are treating the wrong address") this is the second address the protocol does not treat: the electoral formula.
The instrument is captured first — the Hungary/Turkey/India table in 029.10: the autocrat arrives in the trough of trust (19 % in 2009), and after him trust in parliament "rises". Polled trust stops being a bell as soon as the parliament changes owner; the dividend does not.
The thermostat is not blind, it lags. The first draft of this section said "the thermostat is blind to an autocrat with a good economy". The architect's correction: it is blind to freedom, not to the price of the bribe, and for the autocrat that is a trap. The 2022 handout — 6.2 % of GDP out of deficit; then by the textbook: inflation in 2023 — 17.6 %, nominal wages +14 %, the real median −3 %. On the 049 scale that is one or two stars on the phones of the very pensioners and families who received €430 and €850 in February 2022. They do not read opposition papers, but everyone reads the number on the screen — and they will ask not the opposition but him: a year ago you gave, now you took away. The opposition will only name where he lied, and it will not keep silent during growth either. A pre-election handout becomes a debt repaid by the next dividend — it falls by exactly the inflation the handout caused; polled trust the autocrat captures (029.10), the formula he does not (049 §9). Kydland–Prescott (036c.3) in hardware: the price of the lie arrives in a cycle, not in ten years. And the incentive for an autocrat to grow the median is a benefit, not a flaw: Singapore is not a failure of the thermostat but its honest work (034 §80.12); the protocol's measure of freedom is different — double entry and the count.
Orbán's spiral: the autocrat as a state, not a type (015c §7a). In 1989 Orbán is a liberal, the speech at the reburial of Imre Nagy, Fidesz the "Alliance of Young Democrats". In 1998–2002 — prime minister inside a democracy; in 2002 he loses to the socialists running on a "welfare change of regime" (+50 % for public-sector workers, pensions). Lesson one: the honest man is beaten by a handout. In 2006 Fidesz itself already promises a 14th-month pension — and still loses to Gyurcsány, who then admits: "we lied morning, noon and night". Lesson two: the lie is not punished but rewarded. In 2010 — two-thirds, and from then on the price of every next step is zero while the return is confirmed by the ballot box. He did not invent the method: the system showed him twice that the method works, and once that its absence loses. The spiral: a suggestible majority → cunning rewarded → no punishment (13b.3: without punishment defection is the dominant strategy) → confidence that cunning will win next time too → control weakens, because there is nobody left to check: the active are burned and in the kitchen (029.10). Each turn is cheaper than the last — the spiral of lies of 051 §30, lived by one man over twenty years. The conclusion is stronger than "protection from the autocrat": the autocrat is produced by an electorate that pays for manipulation. Remove the payment and the state does not set in; Orbán-2002 with an electorate that does not reward handouts remains who he was.
"A type from the start" is not an objection. For Levitsky Orbán is "an illiberal since the 1990s". Granted. Then the existing system — the one the protocol is compared with — knew it for thirty years and did nothing: did not eject him, did not correct him, did not stop him. And this is not a Hungarian weakness; it is the limit of external control as such. The European Parliament triggered Article 7 in September 2018 — in eight years the Council never voted; Fidesz left the EPP in March 2021 of its own accord, before expulsion; the conditionality mechanism froze €6.3bn in December 2022 — eight months after the election; in December 2023 the Commission unfroze €10.2bn the day before the Ukraine summit where his vote was needed. The hardest instrument the EU has did not change a single ballot in Hungary — but it did fuel the spiral: "Brussels is punishing us" became a personal stake of fear for the very people who were being paid. External control did not stop the state; it fed it. The one lever nobody pulled is the payer: the electorate that pays for manipulation. Type or state is irrelevant to the protocol; what matters is who pays.
"A vaccine, not a cure" — and why that is not so. The first draft left Levitsky one argument: the protocol must be in place before the state sets in, and afterwards four doors are shut. Legal: Art. 8(3) of the 2012 Fundamental Law bans referendums on the budget and taxes. Counting: the count and the media belong to the autocrat. Monetary: the budget must pay, and the budget is his. Human: the majority sleeps, the active are burned. The repository answers all four, and the answers were written before this section. The legal door is removed by 035 (Route B, where Hungary is named directly): no law is needed, only a question — one MP and a private referendum, and the power has four responses, all losing; even in Hungary citizens cannot be forbidden to learn each other's opinion. Orbán has moreover legitimised the form himself: the "national consultations" — postal questionnaires to millions since 2010, whose results he declares the will of the people; the protocol's question is asked in his own form, and he cannot call it illegitimate. The counting door is removed by 056d §7b: a letter in the box under a paper hand count, the last move belongs to the voter — Hungary's ballot is paper. The monetary door is turned over by 056d §5: the autocrat deliberately keeps his voter poor, and the dividend hits those he regarded as his property — the 13th pension named the price. The human door is removed by 057b (a coalition for the right to ask cuts across "for/against" and cannot be declared enemies) and — above all — by 056f: the question is non-political. The burned sit in the kitchen, but in the kitchen the protocol can be discussed and even campaigned for without risk: it is a conversation not against the power but about money and a right, and Orbán's supporter keeps A under it. By Linz (040i) authoritarianism lives by demobilisation and permits precisely the non-political conversation — the protocol enters it as talk about money and leaves it as a demand. What remains for the carrier with a channel (041) is not to build a movement but to say the question aloud.
How it ended: April 2026 (entered 02.10.2026). The section above was written on 28 September and omits the main thing: five months earlier Orbán lost. The omission is the assistant's. On 12 April 2026 Péter Magyar's Tisza won 53.18 % of the vote, Fidesz 38.61 %; turnout was 78.99 % against 69.59 % in 2022. Orbán conceded the same night, and on 9 May parliament elected Magyar prime minister. Four conclusions.
- The thermostat trap worked. The handout of 2022 turned into inflation and a fall in real incomes, and the one who handed it out was held to account — as predicted above.
- The spring released. Plus 9.4 points of turnout — a fourth case beside the three in 011b §11b.2c, and the most recent. The architect holds that the European Union helped remove Orbán; the frozen money was real, but the EU's role does not show in the election results, while the turnout does.
- The European Union's role: words did not work, money did. The assistant first recorded that the EU's role does not show in the election results. The architect: "he could not introduce certain norms in Hungary because of the European Union, and ten billion is no small sum for Hungary either; look in the press at how the EU's struggle with Orbán went". The role shows in three places. Money: since 2022 about 35 billion euros were frozen, of the order of 10 % of GDP, with about 19 billion still frozen at the election; 1.04 billion was lost for good at the end of 2024; in ten months of 2025 the country, for the first time since 2004, paid more into the EU budget than it received; investment fell for about six years. Norms: the law "on the transparency of public life" was postponed in June 2025 after the Commission demanded its withdrawal; the EU court in June 2024 imposed a fine of 200 million euros and a million a day for breaching asylum rules, and in April 2026 held the "child protection" law a breach of the Union's basic values. The campaign: in analysts' judgement Magyar won largely on the promise to bring this money back. This refines the paragraph above on the limit of external control: the Article 7 procedure opened in 2018 never reached a vote in eight years, while the freeze begun in December 2022 worked within one electoral cycle. The European Union did to a country what the protocol does inside a country: it turned the price of bad government into a number visible to everyone. People voted on many things at once — prices, corruption, hospitals; the freeze was not the cause of the defeat but what made the price visible.
- The machine is intact; its owner changed. With 53 % of the vote Tisza took 141 seats of 199 — more than two-thirds. The seat formula that gave Orbán a constitutional majority in 2010 gave it to the next holder: that is "the door the protocol does not close".
- The measure of what was done is how much had to be undone. The architect: "what Orbán did is easily seen from how much has already been repealed and redone after him". In the first months: a constitutional amendment setting an eight-year limit for the prime minister, adopted retroactively; restoration of the rules on the Constitutional Court in force before 2012; abolition of the Sovereignty Protection Office; early termination of the president's term. After that the first 10 billion euros were unfrozen. All of it was done with the same instrument — two-thirds and amendments to the Fundamental Law — that Orbán used.
The price of the existing system's self-correction is sixteen years, and the disease is not cured: the spiral was reversed while elections were still being held, and the machine that drives it remained in place.
Weak point. After the autocrat the protocol works as a question, not as a payment: a weapon of delegitimation from day one, an institution with money on the account only with a change of regime; until then nobody will pay a single forint, and that is the one thing Levitsky keeps. The thermostat's trap has a lag by construction: under 049 the median is a three-year CPI-adjusted average, and the 2023 dip is smeared across three years ("stable, three stars" instead of one); the speed of the trap and protection from tampering pull in opposite directions. A possible resolution — leave the payment base untouched and show in the SMS, next to the three-year base, the one-year change of the median — but the charter (049) is not to be amended for it yet: the architect's decision of 28.09.2026 is that how to display and compute this clearly and accurately for people must be worked out in detail by economists and sociologists; it is an open question for handover, not an amendment. The architect's addition of 30.09.2026: show both comparisons — with one's own past and with the neighbours; "both can be measured, it is not hard". A comparison with oneself alone misleads: "a country that stands still degrades compared with its neighbours", and the star scale of 049, where growth within ±5 % is called "stable", rates falling behind as the norm. And the "abstract fear + need" stratum is split 5/8 by the German proportion (055c §6.3), while the Hungarian rural voter is poorer and more budget-dependent — so more likely to take, and the 42–51 % estimate for Fidesz is overstated, not understated. 🟡
Verdict (reconstruction)¶
"For twenty years we explained that democracies die from the weakness of norms. This mechanism is the first attempt to replace a norm that cannot be held with an incentive that cannot be circumvented. We are not sure it will work. But we are forced to admit: all our prescriptions are addressed to the conscience of elites — this one alone is addressed to their calculator. History votes for the calculator."
Related: 034 (Fraenkel) · 029.4b (the autocrat already pays) · 029.10 (the captured instrument) · 038 (self-verification) · 048b (the economics of dumbing-down) · 048c (the science of the rational killing of a nation) · §9 · §13.9 · 13b.3 · 056c