Part VI: The Crisis

When every governance model began failing — simultaneously, and for the same reason

The twentieth century's stress tests produced survivors. Democracies won the Cold War. The welfare state delivered unprecedented prosperity. International institutions kept great-power peace for seventy years. Then the twenty-first century asked a question none of them were designed to answer: what happens when the feedback loops that made governance work — elections, markets, institutional checks, international cooperation — stop carrying reliable signal? What follows is a three-chapter diagnosis: democracy failing from within, autocracy failing from its own logic, and problems that have outgrown every governance system on earth.

Chapter 18: The Democratic Recession

In 2014, two political scientists at Princeton and Northwestern published a study that should have detonated like a bomb. Martin Gilens and Benjamin Page analyzed 1,779 policy issues over two decades — a wide sample of questions that had come before the American government — and measured whose preferences actually determined what became law. Their finding was stark: what average citizens wanted had essentially no independent effect on what became law. As the preferences of economic elites shifted, policy followed. When elites and ordinary citizens disagreed, elites won. Virtually every time.

The study did not use the word "oligarchy." Journalists did. Gilens and Page, being careful scholars, noted that the preferences of elites and ordinary citizens are highly correlated on most issues — they agree about most things — and the divergence matters primarily on contested questions of regulation and redistribution. Critics challenged the statistical methods. Others argued the model explained too little of the variation. The debate was lively, collegial, and conducted entirely within academic journals.

Meanwhile, outside those journals, something was happening that no study could capture in a regression table. Across the democratic world — not just in fragile young democracies, but in the oldest and most established ones — the system was hollowing out.

Freedom House, the organization that has tracked political freedom globally since the 1970s, began recording a pattern in 2006 that has not reversed. Every year since — nineteen consecutive years — more countries have experienced democratic decline than improvement. In 2024, sixty countries saw democratic declines while only thirty-four improved. The declines affected more than forty percent of the global population. In over forty percent of countries that held elections that year, candidates were targeted with assassination attempts, polling places were attacked, or post-election protests were met with disproportionate force.

And it is not only in the physical world. Internet freedom has declined for over fourteen consecutive years. The digital commons — once imagined as democracy's amplifier — has become "more controlled and manipulated today than ever before." The information environment that democracy depends on has been degraded at the same time as the institutions that process it.

These are not the numbers of a world where democracy is under siege from without. They are the numbers of a world where democracy is failing from within.

The geographic sweep tells the story. Hungary, a European Union member state, has been systematically dismantled as a liberal democracy since Viktor Orbán returned to power in 2010. His methods were not revolutionary — they were bureaucratic. He expanded the Constitutional Court from eleven to fifteen members and packed it with loyalists. Media outlets were captured through state-aligned ownership until the information environment was transformed. Foreign-funded NGOs faced new restrictions, labeling organizations that aided asylum seekers as "foreign agents." He did it all through legislation, through legal processes, through the institutions of democracy itself. And when he was asked what he was building, he answered with disarming honesty: an "illiberal state."

Poland followed the template. Jarosław Kaczyński, leader of the Law and Justice party, made no secret of his admiration, declaring in 2011: "The day will come when we will have Budapest in Warsaw." Between 2015 and 2023, PiS packed courts and constrained media independence, adapting Orbán's playbook for Polish conditions. The EU watched, deliberated, issued stern statements — and largely failed to act. Its governance architecture, designed for cooperation among democratic states, had no effective mechanism for dealing with a member state that was democratically electing its way out of democracy.

The EU's response — or lack of it — is itself a governance lesson. The Rule of Law Mechanism, the EU's primary tool for addressing democratic backsliding in member states, was designed to capture dramatic violations: a coup, a constitutional suspension, a blatant power grab. It could not capture what actually happened in Hungary — the accumulated effect of many small legislative changes, each individually defensible, collectively devastating. Orbán remained a "reliable partner" at the EU level while dismantling democracy at the national level. The toolbox was designed for external threats, not internal corrosion. The democratic rot proceeded through democratic procedures.

This was the pattern that Steven Levitsky and Daniel Ziblatt, two Harvard political scientists, described in their 2018 book How Democracies Die. The title was carefully chosen. Modern democracies, they argued, no longer die through military coups or revolutionary seizures. They die through elected governments gradually subverting the institutions that constrain them. Democracy no longer ends with a bang but with a whimper.

Levitsky and Ziblatt identified four behavioral warning signs: rejection of democratic rules, denial of opponents' legitimacy, toleration or encouragement of violence, and readiness to curtail civil liberties. But their deepest insight was about something harder to measure. Democracies, they argued, depend on two unwritten norms — mutual toleration (treating political opponents as legitimate competitors, not enemies) and institutional forbearance (not exercising every available legal power to its maximum extent). These are norms, not laws. No constitution can mandate them. And both were eroding.

India tells this story at a scale that dwarfs every other example.

The world's largest democracy — over 1.4 billion people, elections conducted across a subcontinent of staggering linguistic, religious, and ethnic diversity — has been reclassified. Freedom House downgraded India from "Free" to "Partly Free" in 2021. The V-Dem Institute at the University of Gothenburg went further, classifying India under Prime Minister Narendra Modi as an "electoral autocracy" — a term unthinkable for a country that has held elections continuously since 1947.

The mechanisms are by now familiar, but in India they operated at a scale and speed that revealed how thin the institutional membrane protecting democracy can be. Key policies — demonetisation, the abrogation of Kashmir's special status, the Citizenship Amendment Act — were made by a small circle around the Prime Minister, often bypassing cabinet deliberation entirely. "Referee institutions" — the judiciary, the Election Commission, investigative agencies — weakened through what scholars describe as a combination of coercion, patronage, and institutional neglect. The Central Bureau of Investigation and Enforcement Directorate were deployed against opposition politicians with a selectivity that strained any claim of impartial justice. Press freedom deteriorated until Reporters Sans Frontières ranked India 161st globally. Internet shutdowns became routine — India imposed more internet blackouts than any other democracy, often during moments of political protest. The National Register of Citizens in Assam, intended to identify undocumented migrants, left nearly two million people — many of them lifelong residents — in bureaucratic limbo, uncertain of their citizenship in the country of their birth.

What made India's case distinctive was not the erosion itself but what it revealed about scale. A nation of 1.4 billion people, with twenty-two official languages, extraordinary religious and ethnic diversity, and a democratic tradition stretching back to 1947, was being reshaped through the same playbook that had transformed a European nation of ten million. The mechanisms were portable. The institutional defenses were not.

But India also demonstrated something the pessimists had not predicted. In the 2024 general election, the BJP lost its parliamentary majority for the first time since 2014. The institutional capacity to hold elections and express dissent — however weakened — still functioned. For the first time in years, V-Dem recorded no further democratic deterioration in India. Whether institutional repair will follow electoral correction remains an open question. But the correction happened. The feedback loop, attenuated and degraded, still carried a signal.

The United States told a grimmer version of the same story. In 2025, the Century Foundation launched its "Democracy Meter." Its first reading registered an "authoritarian turn" — no improvements recorded in any category. In the country that had spent two centuries positioning itself as democracy's exemplar, the institutional architecture was straining under pressures its framers had not imagined and its modern stewards had not adequately addressed.

What is happening? Why is the democratic model failing across such different societies, cultures, and institutional designs simultaneously?

The structural explanations converge on a single insight: the mechanisms that democracies use to detect the preferences of their citizens have degraded to the point where governance responds to something other than public need.

Consider the mechanics. In the United States, gerrymandering — the drawing of electoral districts to predetermine outcomes — has reached mathematical precision. In Wisconsin's 2012 election, using maps drawn by the new Republican legislature the previous year, Republicans won 60 of the State Assembly's 99 seats despite Democrats receiving more total votes statewide. A 2020 study found that gerrymandering doesn't just distort representation; it suppresses participation — fewer candidates contest disadvantaged districts, those who run have weaker credentials, donors contribute less, and voters disengage. The outcome is predetermined before the ballot is cast. The feedback loop from citizen preference to policy outcome is severed before the ballot is cast.

Layer on money. The connection between campaign finance and gerrymandering operates as a reinforcing cycle: corporate spending can buy a gerrymander, and elected officials who benefit from the resulting distortion have no incentive to reform the system. Gilens and Page's finding — elite preferences driving policy on contested economic issues — is not mysterious when a competitive congressional race costs over ten million dollars and the median U.S. senator's net worth tops one million. The feedback loop doesn't connect citizens to representatives. It connects donors to representatives.

Then add the information environment. The echo-chamber story — algorithms trapping people in bubbles of self-reinforcing belief — turns out to be both simpler and more alarming than the metaphor suggests. When researchers actually switched users from algorithmic to chronological social media feeds, the users' political views barely budged. The bubbles, it turned out, were not the main problem. The main problem was subtler: a media architecture that rewards emotional intensity over factual accuracy, that amplifies outrage because outrage keeps people scrolling, and that degrades the shared factual reality on which democratic deliberation depends. People are not living in separate chambers. They are living in the same room with incompatible maps of what the room contains.

The European Union presents a different failure mode. Here the problem is not polarization or capture — it's opacity. The EU's democratic deficit operates through sheer institutional complexity. The European Commission, which alone can initiate legislation, is not directly elected. Most legislation is concluded through informal negotiations between Parliament, Council, and Commission that operate largely outside public view — a process Chapter 20 will examine in detail. European Parliament elections draw low turnout — voters treat them as second-order national contests rather than genuine EU governance choices. The result is governance that may be technically competent but is democratically hollow: few citizens can explain how it works, which means few citizens can meaningfully contest its outcomes.

Here is the pattern that connects Wisconsin's gerrymandered maps to Brussels' trilogues, Hungary's captured courts to India's weakened institutions, America's donor-driven policy to the EU's technocratic opacity: democratic governance has stopped detecting resonance and started performing ritual.

Elections still happen. Parliaments still convene. Laws still pass. But the feedback loop — the mechanism by which the governed communicate their needs to the governing, and the governing adjust in response — has been severed, degraded, captured, or rendered so complex that it no longer carries useful signal. Democracy performs its forms while losing its function. The ritual continues. The resonance is gone.

And because people can feel the difference — can sense that voting doesn't change things, that their concerns don't register, that the system responds to something other than their needs — trust collapses. Disengagement follows. And into the space created by disengagement, opportunists move: populists who name the disconnect (correctly) while offering solutions that deepen it (catastrophically).

This is not a story about bad actors defeating good institutions. It is a story about institutions designed for eighteenth-century information environments failing in twenty-first-century conditions. Representative democracy was built when communication meant pamphlets and physical assemblies, when feedback moved at the speed of horse and sail, when a legislator could plausibly claim to know their constituents' needs through personal acquaintance. That architecture has not fundamentally changed in two centuries. The world has.

But the story does not end with decline. It also includes response. Citizens' assemblies — groups of randomly selected ordinary people deliberating on complex policy questions — have proliferated across the democratic world. Ireland used them to navigate the politically impossible question of abortion reform. France convened a Convention Citoyenne pour le Climat with 150 randomly selected citizens. New York City began piloting assemblies integrated with participatory budgeting in 2024. The EU itself recognized participatory budgeting as "a pathway to inclusive and transparent governance."

These experiments share a common logic: if elections no longer carry adequate signal between governed and governing, create new channels. Bypass the captured mechanisms. Go directly to citizens. Ask them to deliberate — not just vote, but reason together about trade-offs, hear competing evidence, and reach considered judgments.

Participatory budgeting — the practice of giving citizens direct authority over portions of public spending — tells the story most concretely. It began in Porto Alegre, Brazil, in 1989, when a newly elected Workers' Party government invited residents to decide how to allocate municipal investment. Neighborhoods held assemblies. Delegates were chosen. Priorities were ranked — sewerage in this favela, a school in that district, road paving where the buses couldn't run during the rainy season. The results were measurable: districts that had been ignored for decades received infrastructure for the first time. Studies found significant reductions in infant mortality in municipalities that adopted participatory budgeting, driven by increased spending on sanitation and healthcare in the poorest areas. The feedback loop was restored — not through elections but through direct engagement with the consequences of spending decisions.

From Porto Alegre, the model spread. Over 11,500 municipal processes globally. Seoul dedicated a portion of its city budget to citizen-directed allocation. Paris committed one hundred million euros annually. New York City began piloting assemblies integrated with participatory budgeting in 2024. The amounts are often modest — a fraction of total budgets. But the governance logic is significant: people who decide how money is spent pay attention to what it buys.

Liquid democracy — where citizens can vote directly on issues or delegate their vote to a trusted proxy, issue by issue — has been attempted by parties in Argentina and Australia, though early evidence suggests delegation tends to recreate the same power imbalances it was designed to bypass. The results are mixed. The experiments are young.

What matters is not whether any single innovation succeeds. What matters is the pattern: when a governance system's feedback loops degrade, people begin building new ones. The democratic recession is real, it is deep, and it is global. But so is the response — the search for mechanisms that can restore the connection between the governed and the governing.

The welfare state's golden age showed that governance can work — that the feedback loop between citizen and state, when functioning, produces extraordinary results. The democratic recession shows what happens when that loop degrades. The question is no longer whether democratic governance is the right model. The question is whether it can be repaired before the alternatives — which are already waiting, already growing, already governing billions — make the question irrelevant.

Those alternatives are the subject of the next chapter. They have their own strengths, their own appeal, and their own dilemma — a dilemma that no amount of efficiency can resolve.