Why Democracies Keep Surviving Their Own Obituaries

For a decade the headline has been decline. Strongmen rising, elections questioned, institutions bending. "Democracy" rarely appears in print without "crisis" nearby, and surveys across rich countries show citizens doubting their systems can still deliver. The mood is that of a patient on a falling chart.

Here is the less fashionable observation — which does not make it true, only worth testing: established democracies have looked this fragile before and did not die. Some have a repeated habit of absorbing shocks that, by any reasonable forecast, should have finished them.

The situation

The anxiety is grounded. Several leaders have tested the limits of their offices. Trust in politicians is low. Affective polarisation — treating the other side as dangerous rather than merely wrong — has risen measurably in the United States and parts of Europe. Freedom House, which scores countries on political rights and civil liberties, has recorded more years of global decline than improvement since around 2006.

So the diagnosis of weakness is honest. The open question is what that weakness predicts — bending or breaking. That is where a model helps.

The model, and its status

The writer Nassim Taleb popularised the term antifragile for things that are not merely resistant to stress but improved by it. A fragile object breaks under shock; a robust one endures unchanged; an antifragile one responds by getting stronger — like a muscle growing after being torn under load, or an immune system that learns from an infection it survives.

Be clear about what this is. "Antifragile" is a useful frame, not a tested law; Taleb coined it in a 2012 book of argument, not from a dataset. The empirical question is whether mature democracies actually behave like the muscle rather than the glass. The frame is only worth keeping if the mechanism is real and the record supports it.

The mechanism

Three steps, each checkable.

A threat is visible before it is fatal. Democratic backsliding is usually slow and public: courts packed, press squeezed, norms broken one at a time. Visibility buys opponents time to organise before the game is lost.

The threat manufactures its own opposition. When a leader overreaches, people who agree on nothing else discover a shared interest in stopping him. Business leaders, judges, opposition parties, parts of the civil service and ordinary voters form coalitions that would never otherwise exist. The attack builds the alliance that defeats it.

The defeat gets encoded. After the crisis, the near-miss becomes a lesson. Loopholes close, laws tighten, and — most powerfully — a norm hardens in public memory: we do not do that here. The next aspirant faces a population that recognises the playbook because it watched the last one run it.

The evidence — named

The clearest encoded lesson is postwar Germany's streitbare Demokratie, "militant democracy": a constitution that lets the state ban anti-democratic parties, written precisely because Weimar's openness had been used to destroy it. That is a near-death experience turned into law.

More recent cases cut both ways, which is the point. In the United States in 2020–21, courts (many with judges appointed by the losing side) rejected dozens of election challenges, state officials certified results, and a mob was repelled — then, crucially, several states tightened certification rules afterward. In Brazil in 2023, after rioters stormed the capital, the institutions held and the courts moved hard. In Poland, voters removed a backsliding government at the ballot box in 2023, though reversing its capture of the courts is proving slow and contested.

Against these: Hungary, a democracy that bent and then largely broke, its elections still held but no longer fair. Turkey went the same way. Both were younger democracies with thinner institutions. The immune response is not automatic.

This reframes the scoreboard. The indices showing more decline than improvement are measuring real erosion — but erosion and collapse differ. Many democracies have bent in the past decade. Among the long-established ones, few have actually broken.

The strongest counter-case

The honest objection is that this time the stressors differ in kind, not degree.

First, the information environment. The immune response runs on shared recognition — we all saw him cross the line. When half the population is told no line was crossed, that recognition may never form.

Second, antifragility has a ceiling. A muscle grows under load and tears under too much; a system strengthened by ten near-misses can be destroyed by the eleventh. Resilience so far does not prove resilience next time.

Third, the model travels badly. It fits Denmark far better than a twenty-year-old democracy with weak courts — exactly the pattern Hungary and Turkey show. Where institutions and public memory are thin, the mechanism is unreliable.

What is fact: established democracies have repeatedly survived threats that looked terminal, often producing defensive reforms. What is interpretation: that this is a durable property rather than luck and favourable conditions. What is unknown: whether a fractured, high-speed information landscape has quietly disabled the recognition step the whole thing depends on.

Where else this applies

The shape recurs. The cleanest case is literal: a vaccine is controlled exposure that strengthens the body against the real thing. Financial systems that survive a crash often emerge with better capital rules and fewer of the weaknesses that caused it. In a career, the person who has survived a public failure and rebuilt often has judgment the untested high-flyer lacks — not despite the failure but because of it.

The transferable lesson cuts against instinct: we admire the system, institution or person that has never faced a crisis. The antifragility frame says worry about them more, not less. Smoothness can be brittleness that has simply never been tested.

What to watch

Two signals separate the readings.

First, whether cross-cutting coalitions still form under pressure. If opponents who disagree on everything still unite to stop an overreach, the mechanism works. If attacks meet silence or tribal cheering, it has broken.

Second, whether near-misses produce reforms or just exhaustion. A healthy system turns a crisis into a tightened rule; a declining one turns it into a new normal nobody reverses.

A falsifiable expectation, offered with low confidence because the base rate is small and the frame untested: among democracies with several decades of continuous free elections and independent courts, most that suffer a serious backsliding episode in the next ten years will recover rather than tip into authoritarian rule. Among democracies younger than thirty, the rate will be markedly lower. If an old, thick-institutioned democracy tips outright — or if a wave of them bends without any coalition forming to resist — treat the model as failing. The point is not the number; it is that the frame tells you precisely where, and under what conditions, it should break.

Distilled from The Economist — International