The Scale of the Collapse
Twenty years of democratic recession
Global freedom has declined for 20 consecutive years — ✓ Established Fact — the longest sustained democratic recession in modern history [1]. In 2025, 54 countries experienced deterioration in their political rights and civil liberties, while only 35 registered improvements [1]. The numbers are unambiguous. Democracy is contracting at a pace and scale that demands structural explanation, not isolated case studies.
The V-Dem Institute's 2026 Democracy Report — its tenth annual edition — delivers the starkest assessment yet: democracy for the average person on Earth has fallen back to 1978 levels [1]. Autocracies now outnumber democracies globally for the first time in two decades — 92 autocracies versus 87 democracies, by V-Dem's Regimes of the World classification ✓ Established Fact [1]. The reversal is not marginal. It represents a fundamental shift in the global distribution of political power, undoing gains that took decades to achieve.
The population figures are even more alarming. Approximately 74 per cent of the world's population — roughly 6 billion people — now lives under autocratic governance [1]. Only 21 per cent of the global population lives in countries rated Free by Freedom House, down from 46 per cent just two decades ago ✓ Established Fact [2]. Electoral autocracy has become the most populous regime type, housing nearly half the world's population — 46 per cent, or 3.8 billion people [1]. Liberal democracies — the gold standard of governance — now shelter just 7 per cent of humanity, approximately 600 million people [1].
The current wave of autocratisation is not the first in modern history, but it is distinct in character. Samuel Huntington's concept of the "third wave" of democratisation — the surge of democratic transitions that began with Portugal's Carnation Revolution in 1974 and accelerated through the fall of the Berlin Wall and the collapse of the Soviet Union — shaped an entire generation of political science [7]. But that wave has now been met by what V-Dem researchers term the "third wave of autocratisation," encompassing 105 autocratisation episodes in 75 countries over the past 25 years ✓ Established Fact [4].
International IDEA's Global State of Democracy 2025 report confirms the structural nature of this decline. In 2024, 94 countries — representing 54 per cent of all nations assessed — suffered a decline in at least one factor of democratic performance compared with their own performance five years earlier [3]. This was the ninth consecutive year in which more countries declined than improved — the longest consecutive fall since IDEA's records began in 1975 [3]. Representation scores collapsed to their worst level in over 20 years, even as an unprecedented 74 national elections were held in 2024 [3]. The problem is not a lack of elections. It is what happens around them, before them, and after them.
Freedom House's 2026 report identifies the drivers with precision: military coups, violence against peaceful protesters, and systematic efforts to weaken constitutional safeguards [2]. Media freedom, freedom of personal expression, and the right to due process have registered the sharpest downgrades globally [2]. The decline is not confined to fragile states or post-conflict zones. It has reached the institutional heartlands of liberal democracy — including, most notably, the United States, the United Kingdom, and Italy [1].
This is not a temporary dip or a cyclical correction. Twenty consecutive years of decline across every major democracy index — Freedom House, V-Dem, International IDEA, the Economist Intelligence Unit — constitutes a structural transformation of the global political order. The post-Cold War assumption that democracy would inevitably spread has been decisively falsified. The question is no longer whether democratic recession is real. It is whether it is reversible.
Six of the ten new autocratising countries identified in the 2026 V-Dem report are in Europe and North America — including large and influential countries such as Italy, the United Kingdom, and the United States [1]. This marks a decisive shift. For the first two decades of democratic recession, the decline was concentrated in the developing world, in countries with shorter democratic histories and weaker institutions. That is no longer the case. Autocratisation has reached the countries that built the post-war liberal democratic order.
The 44 countries currently experiencing autocratisation are home to 41 per cent of the world's population — a record high for the current wave ◈ Strong Evidence [1]. Meanwhile, only 18 countries — 10 per cent of the global total — are currently democratising, and they account for just 5 per cent of the world's population [1]. The asymmetry is stark. For every country moving toward democracy, more than two are moving away from it. The democratic recession is accelerating, not stabilising.
The Autocrat's Playbook
How democracies are dismantled from within
Modern democratic backsliding rarely begins with tanks in the streets — ◈ Strong Evidence. It begins at the ballot box, with elected leaders who gradually subvert the institutions that brought them to power [7]. The mechanisms are now well documented: executive overreach, judicial capture, media control, electoral manipulation, and the systematic suppression of civil society. The playbook is remarkably consistent across continents.
Steven Levitsky and Daniel Ziblatt, in their landmark 2018 study How Democracies Die, identified the defining feature of contemporary authoritarianism: it operates through legal and constitutional mechanisms rather than overt force [7]. The would-be autocrat wins an election, then uses democratic institutions against themselves — packing courts, rewriting constitutions, capturing regulatory bodies, and criminalising opposition under the guise of anti-corruption or national security. The process is incremental, often spanning years, and each individual step can be defended as technically legal.
Research by Levitsky and Ziblatt demonstrates that since the Cold War, most democratic breakdowns have been caused not by generals and soldiers but by elected governments themselves [7]. The process involves the gradual erosion of what they term democracy's "soft guardrails" — mutual toleration (accepting opponents as legitimate) and institutional forbearance (restraint in the exercise of power). Once these norms collapse, formal institutions follow.
The V-Dem Institute's 25-year dataset reveals the most common targets. Freedom of expression shows the most drastic global decline and is the most frequently attacked dimension of democracy by autocratising leaders [4]. Government censorship of the media now affects 32 of the 44 currently autocratising countries, while repression of civil society affects 30 [4]. Self-censorship among journalists when reporting on politically sensitive issues is becoming more common in 32 countries, including India, Turkey, and Hungary ✓ Established Fact [4].
The pattern follows a predictable sequence. First, the executive weakens judicial independence by packing courts with loyalists, altering appointment procedures, or expanding court size. Second, independent media is brought under control through regulatory pressure, ownership consolidation, or direct acquisition by government-aligned entities. Third, civil society organisations are constrained through legislation restricting foreign funding, surveillance, or criminal prosecution of activists. Fourth, the electoral playing field is tilted through gerrymandering, voter suppression, or manipulation of electoral management bodies.
What makes this playbook so effective is its incrementalism. No single step constitutes a dramatic break with democratic norms. A court-packing scheme can be framed as judicial reform. Media consolidation can be presented as market efficiency. Restrictions on NGO funding can be justified as transparency measures. Each action is individually defensible. Collectively, they constitute the architecture of authoritarian control.
The 2025 Global Democracy Conference at the University of Notre Dame examined how executives — from Hungary to Venezuela — have systematically undermined independent institutions and cleared the path toward authoritarianism [7]. Anti-democratic executives have progressively packed the judiciary, purged the civil service, undermined electoral management bodies, silenced independent media, prosecuted dissidents, restricted non-governmental organisations, regulated the business sector in favour of cronies, and politicised the security forces. The pattern is global and consistent.
The most dangerous feature of modern autocratisation is its legality. Unlike Cold War-era coups, which violated constitutions openly, contemporary democratic erosion operates through constitutional amendments, legislative supermajorities, and judicial reinterpretation. In Mexico, the 2024 judicial reform — electing all judges by popular vote — was passed through legal channels. In El Salvador, Bukele's unconstitutional re-election was ratified by a compliant supreme court. The law itself becomes the instrument of democratic destruction.
Government intimidation of the opposition during election periods has increased substantially in 21 countries, among them India, Turkey, and Hungary [4]. Electoral manipulation need not involve fraud in the counting of ballots. It operates through control of media access, selective prosecution of opponents, manipulation of registration systems, and structural advantages embedded in redistricting. In Turkey, President Erdoğan maintains the facade of competitive elections while the underlying institutional environment ensures the playing field is permanently tilted [4].
The CIVICUS State of Civil Society Report 2025 identifies crackdowns on civil society organisations — including criminalisation of activists, smear campaigns, surveillance, and prohibition of funding — as both a symptom of and a prelude to democratic decline ◈ Strong Evidence [11]. When civil society is suppressed, the early warning system for democratic erosion is dismantled. Citizens lose access to independent information, organised advocacy, and the institutional capacity to resist executive overreach. The autocrat does not need to abolish democracy. They need only to hollow it out.
The Evidence From Within
Hungary, India, Turkey, and El Salvador
Four case studies illustrate how the autocrat's playbook operates in practice — each in a different institutional context, each following a recognisable pattern, and each offering lessons about the conditions under which democracies fail ✓ Established Fact. The evidence is drawn from court records, internal government documents, international assessments, and investigative journalism.
Hungary — The Laboratory of Illiberal Democracy. Viktor Orbán's Hungary is the most thoroughly documented case of democratic backsliding within the European Union. Since returning to power in 2010 with a constitutional supermajority, Orbán has systematically dismantled every independent institution in Hungary while maintaining the formal structures of electoral democracy [13]. The Cato Institute's detailed analysis documents how the process unfolded: first, the Constitutional Court was packed with loyalists and its powers curtailed through constitutional amendment. Then, the judiciary was restructured to increase political influence over appointments. Media was consolidated through a government-aligned foundation — the Central European Press and Media Foundation — which in 2018 received approximately 500 media outlets transferred by owners close to the Orbán government [13].
The results are measurable. Hungary's Freedom House rating has fallen from near-perfect scores in 2010 to a "Partly Free" classification at 65 out of 100 [13]. Transparency International's Corruption Perceptions Index score dropped 14 points — from 56 to 42 — between 2012 and 2023 ✓ Established Fact [13]. The EU has frozen €6.3 billion of Hungary's cohesion funds and an additional €9.6 billion in Recovery and Resilience Facility grants over rule of law violations [13]. Independent journalists face smear campaigns, are banned from government press conferences, and in some cases have been subjected to Pegasus spyware surveillance [13]. Orbán himself has coined the term "illiberal democracy" — a contradiction in terms that has nonetheless become a model for aspiring autocrats worldwide.
Democratic backsliding today begins at the ballot box. The electoral road to breakdown is dangerously deceptive. Because there is no single moment — no coup, no declaration of martial law — nothing may set off society's alarm bells.
— Steven Levitsky & Daniel Ziblatt, How Democracies Die, 2018India — The World's Largest Electoral Autocracy. India's democratic decline is arguably the most consequential in the current wave, given the country's population of 1.4 billion. The V-Dem Institute has classified India as an "electoral autocracy" since 2017 — a designation that the Modi government has vigorously contested but that multiple independent assessments confirm [14]. India now ranks 100th of 179 countries on V-Dem's Liberal Democracy Index, and Freedom House downgraded it from Free to Partly Free in 2021 [14].
The mechanisms are familiar: harassment of journalists critical of the government, attacks on civil society and the opposition, and government intimidation during election periods [4]. Self-censorship among journalists has become widespread. The 2024 elections — in which the BJP was forced into coalition governance after losing its outright majority — represented a partial correction, and 2024 was the first year since 2008 with no further democratic deterioration in India [4]. But the structural damage to institutions accumulated over a decade may take far longer to repair.
Turkey — The Facade of Competition. Turkey's trajectory under Recep Tayyip Erdoğan demonstrates how democratic institutions can be hollowed out while maintaining the appearance of competitive elections. Following the failed 2016 coup attempt, Erdoğan transformed Turkey from a parliamentary system to an executive presidency, concentrating enormous power in the office of the president while systematically weakening judicial independence, press freedom, and civil society. The arrest of Istanbul Mayor Ekrem İmamoğlu in March 2025 — days before he was expected to be selected as the opposition's presidential candidate — illustrates the pattern: use the legal system to neutralise political rivals while claiming to uphold the rule of law.
El Salvador — The Popular Autocrat. Nayib Bukele represents a particularly challenging variant of democratic backsliding: the autocrat with genuine popular support. His crackdown on gang violence — involving a permanent state of exception extended 36 times by the legislature, with over 85,000 suspected gang members arrested — has won overwhelming domestic approval [12]. But the democratic cost is severe. El Salvador now has the highest incarceration rate in the world, at 1.7 per cent of its population ✓ Established Fact [12]. The country fell 61 places in the World Press Freedom Index since 2019 [12]. In July 2025, the congress removed presidential term limits entirely, formalising the concentration of power that had already occurred in practice.
Since 2010, Orbán's Fidesz party has rewritten the constitution, packed the Constitutional Court, consolidated 500 media outlets under a government-aligned foundation, and used Pegasus spyware against independent journalists [13]. The EU has responded by freezing €15.9 billion in funds, but the Article 7 procedure — designed to sanction rule of law violations — remains stalled because it requires unanimity, which Hungary's allies can block.
The four cases share common features: executive aggrandisement, judicial capture, media control, and the use of legal mechanisms to suppress opposition. They differ in context — Hungary operates within the EU, India within the world's largest electorate, Turkey within NATO, and El Salvador within a region historically prone to authoritarianism. But the convergence of methods is striking. Aspiring autocrats are learning from one another, adapting the playbook to local conditions while following the same structural logic.
The Human Cost
When democratic guardrails fail
Democratic backsliding is not an abstract institutional phenomenon. It translates directly into human suffering — imprisoned journalists, silenced activists, displaced populations, and citizens stripped of fundamental rights ✓ Established Fact. The data on press freedom alone reveals the scale of the crisis [6].
Reporters Without Borders' 2025 World Press Freedom Index recorded the lowest global score in the index's history — a global average of 55, classified for the first time as a "difficult situation" [6]. The human toll is staggering: as of 1 December 2025, 67 reporters had been killed, 503 detained, 135 were missing, and 20 were held hostage worldwide ✓ Established Fact [6]. Nearly half — 43 per cent — of the journalists killed were in Gaza. China remains the world's largest prison for journalists, with 113 media professionals detained under Xi Jinping's regime, followed by Russia with 48, including 26 Ukrainian journalists [6].
Over half the world's population now lives in what RSF classifies as press freedom "red zones" — areas where practising journalism carries significant personal risk [6]. Fewer than 8 per cent of the global population lives in places where the situation is classified as "good" or "satisfactory" [6]. The collapse of press freedom is not merely a symptom of democratic decline — it is an accelerant. When independent journalism is suppressed, citizens lose access to the information they need to hold power accountable, and the feedback mechanisms that sustain democratic governance are severed.
The suppression of civil society compounds the damage. CIVICUS documents a global pattern of crackdowns on civil society organisations that follows a consistent sequence: first, legislation restricting foreign funding for NGOs; second, surveillance and infiltration of activist networks; third, criminalisation of protest and dissent; fourth, smear campaigns designed to delegitimise civil society actors in the public eye [11]. These crackdowns are both a symptom of democratic decline and a prelude to its acceleration — the dismantling of the institutional infrastructure through which citizens organise, advocate, and resist.
The imprisonment of 503 journalists worldwide is not merely a statistic. Each detained journalist represents a story that will never be told, an investigation that will never be completed, and a community that will never receive the information it needs to govern itself. The suppression of press freedom is not a side effect of autocratisation. It is the mechanism through which autocrats consolidate control — by severing the informational arteries of democratic governance.
The United States — historically a global standard-bearer for press freedom — has not been immune. RSF reported that the US saw the sharpest drop in press freedom in the Americas in 2025, falling from a "Low Restrictions" classification to "With Restrictions" under President Trump's second administration [6]. The country now ranks 57th globally on the Press Freedom Index. Freedom House's 2026 report specifically cites "a multiyear rise in threats and reprisals for nonviolent speech" as a driver of the US's score decline to 81 out of 100 [10].
The human cost extends beyond the quantifiable. In countries experiencing democratic backsliding, citizens report declining trust in institutions, increasing self-censorship, and a pervasive sense that the rules of the game have changed. The V-Dem data on freedom of expression — deteriorating in 44 countries, the highest number ever recorded — captures only the institutional dimension [4]. The psychological dimension — the chilling effect on public discourse, the erosion of social trust, the normalisation of authoritarian behaviour — is harder to measure but no less real.
El Salvador's state of exception offers a concentrated illustration. With 85,000 arrested and 1.7 per cent of the population incarcerated, the human rights cost of Bukele's security crackdown has been enormous [12]. The Organisation of American States Electoral Observation Mission flagged "deficiencies and bad practices that affected aspects of fairness and transparency" in the 2024 elections — held under the permanent state of exception. Due process has effectively been suspended for an indeterminate period. The popular support for the crackdown does not change its character: democratic rights are not subject to majority approval. They exist precisely to protect minorities and individuals from the exercise of unchecked power.
Country by Country
A comparative anatomy of decline and recovery
Democratic backsliding does not follow a single template. The trajectories of individual countries reveal the diversity of pathways into — and, in some cases, out of — autocratisation ◈ Strong Evidence. A comparative analysis of key cases illuminates both the structural vulnerabilities that enable backsliding and the conditions under which reversal becomes possible [5].
The United States represents the most consequential case of democratic backsliding in the current wave. V-Dem's 2026 report classifies the US decline as "unprecedented" in its speed and magnitude — a 24 per cent decline on the Liberal Democracy Index in a single year, dropping the country from 20th to 51st globally ✓ Established Fact [1]. The Century Foundation's Democracy Meter rates the US at 57 out of 100 — a 28 per cent drop in one year [15]. Freedom House recorded the sharpest decline among Free nations, with the US score falling to 81 out of 100 — its lowest since scoring began in 2002 [10].
The core diagnosis is consistent across all three major indices: the executive has aggrandised power beyond what a liberal democracy can sustain, supported by a highly partisan Supreme Court and with Congress unwilling to intervene [15]. Freedom of expression, rule of law, and checks and balances have all deteriorated sharply. The US case demonstrates that no democracy — regardless of its age, wealth, or institutional depth — is immune to backsliding when political elites abandon the norms of mutual toleration and institutional forbearance that Levitsky and Ziblatt identified as democracy's critical "soft guardrails" [7].
Poland offers the most encouraging recovery story in Europe. Under the Law and Justice party (PiS) from 2015 to 2023, Poland experienced systematic erosion of judicial independence, media freedom, and civil society space — severe enough to trigger the EU's Article 7 proceedings in December 2017 [5]. The 2023 election of a coalition government under Donald Tusk began a process of democratic restoration. Poland improved in six democratic performance factors between 2023 and 2024, and in May 2024, the EU formally closed the Article 7 procedure, concluding there was no longer a clear risk of a serious breach of the rule of law [5].
However, Poland's recovery also reveals the limits of electoral alternation. PiS's eight years in power allowed it to embed loyalists in the Constitutional Tribunal and the National Council of the Judiciary, and these appointees have actively blocked the new government's attempts to restore civil liberties [5]. Democratic recovery, even in the best-case scenario, is slower and harder than democratic erosion. The damage done to institutions persists beyond the government that inflicted it.
Brazil represents another successful reversal. After years of democratic deterioration under Jair Bolsonaro — including executive aggrandisement, attacks on the judiciary, and media restrictions — the election of Luiz Inácio Lula da Silva in 2022 initiated a recovery. Brazil recorded advances in 10 democratic performance factors between 2022 and 2024 ✓ Established Fact [5]. International IDEA notes that targeted countering of disinformation regarding elections was a key factor in Brazil's democratic reversal [3].
Tunisia stands as a cautionary tale. The sole democratic success of the Arab Spring collapsed in 2021 when President Kais Saied suspended parliament, and by 2022 had passed a new constitution granting himself virtually unchecked power. The legislative branch was rendered powerless. Opposition figures, particularly those associated with the Ennahda Party, have been arrested on broad charges of "conspiracy against state security." Tunisia's trajectory demonstrates that democratic gains can be erased with alarming speed when institutional safeguards are weak and when no external accountability mechanism exists.
South Korea provides the most dramatic example of democratic resilience. On 3 December 2024, President Yoon Suk Yeol declared martial law in a televised address — the first martial law since South Korea's democratisation in 1987 [9]. The legislature convened and voted to lift it within six hours. Citizens mobilised rapidly, drawing on decades of democratic resistance culture. The Constitutional Court unanimously upheld Yoon's impeachment and removal in April 2025 [9]. Korean citizens who mobilised against the martial law decree were nominated for the Nobel Peace Prize. South Korea's case confirms that democratic resilience is not automatic — it depends on the strength of institutions, the mobilisation capacity of civil society, and the willingness of key actors (legislators, judges, military leaders) to uphold constitutional obligations.
When Democracy Fights Back
The conditions for reversal
If the evidence shows that democracies can die, it also shows that they can recover — but only under specific conditions ◈ Strong Evidence. Carnegie's research on democratic recovery identifies four "U-turn" countries that halted and reversed autocratisation before full breakdown: Brazil, Ecuador, Lesotho, and Poland [5]. Understanding what they share — and what distinguishes them from cases like Hungary and Turkey — is critical for any assessment of reversibility.
The Carnegie Endowment's 2025 study on democratic recovery after significant backsliding identifies several common factors in successful reversals [5]. First, electoral alternation — the defeat of the autocratising incumbent through competitive elections — is the most consistent trigger for recovery. In all four U-turn cases, a change of government preceded democratic restoration. Second, civil society mobilisation played a critical role in sustaining pressure on institutions and providing the organisational infrastructure for political opposition. Third, in Brazil specifically, targeted counter-disinformation regarding elections was identified as a key factor in the reversal [3].
Brookings research emphasises the concept of "diagonal accountability" — the mechanism through which civil society organisations use protests, research, advocacy, and information dissemination to keep governments responsible and responsive [8]. This form of accountability, the research argues, is just as important as "horizontal accountability" — the formal checks and balances between branches of government [8]. When horizontal accountability fails — when courts are captured, legislatures are compliant, and regulatory bodies are co-opted — diagonal accountability may be the last line of defence.
Carnegie's study of the four U-turn countries reveals that no single factor is sufficient [5]. Brazil's recovery required both electoral victory and systematic counter-disinformation. Poland's required both a new government and EU institutional pressure. Ecuador and Lesotho's required both political alternation and civic mobilisation. Where one element is missing — civil society in Russia, electoral alternation in Hungary — recovery stalls or fails entirely.
However, the conditions for recovery are narrow. The Carnegie research implicitly identifies the critical threshold: recovery is possible only before democratic breakdown is complete [5]. Once an autocrat has fully captured the judiciary, eliminated independent media, and suppressed civil society, the mechanisms through which recovery occurs — competitive elections, judicial challenges, civic mobilisation — are no longer available. Hungary illustrates this trap: even after a hypothetical electoral defeat of Fidesz, the Constitutional Tribunal packed with Orbán loyalists would remain in place, capable of blocking reform from within the institutions.
South Korea's case offers a counterpoint and a clarification. Recovery there was possible because the autocratic attempt was sudden and dramatic — a martial law declaration — rather than the gradual, incremental erosion characteristic of Hungary or Turkey. The institutions were still intact. The legislature was still functional. The judiciary was still independent. The military chose not to follow illegal orders. Civil society was able to mobilise because the infrastructure for mobilisation had not been dismantled [9]. South Korea's success was not evidence that resilience is easy — it was evidence that resilience requires institutions that have not yet been captured.
The EU's conditionality mechanism offers a model for external pressure, but one with significant limitations. The freezing of €15.9 billion from Hungary has imposed real costs on the Orbán government, but it has not reversed the democratic erosion [13]. The Article 7 procedure — theoretically designed for precisely this purpose — remains structurally blocked because it requires unanimity, and Hungary has allies willing to protect it. The EU's success with Poland — where Article 7 proceedings were closed after the change of government in 2023 — was driven primarily by domestic political change, not by Brussels's institutional leverage.
| Risk Factor | Severity | Assessment |
|---|---|---|
| Judicial Capture Before Electoral Alternation | Once courts are packed with loyalists, new governments face legal obstruction from within the state apparatus. Hungary and Poland both demonstrate this — Poland's recovery is slowed by PiS appointees in the Constitutional Tribunal who block reform. | |
| Media Consolidation Under Autocratic Control | When 500 media outlets are consolidated under a government-aligned foundation (Hungary) or press freedom collapses (El Salvador, -61 places), the information environment necessary for democratic recovery is destroyed. | |
| Civil Society Suppression | CIVICUS data shows crackdowns on CSOs — criminalisation, surveillance, funding restrictions — precede and accelerate democratic decline. Without organised civil society, the mobilisation infrastructure for recovery does not exist. | |
| Constitutional Entrenchment of Executive Power | Tunisia's 2022 constitution and El Salvador's 2025 removal of term limits demonstrate how constitutions can be rewritten to permanently concentrate power. Once embedded in constitutional text, reversal requires extraordinary supermajorities. | |
| Absence of External Accountability | The EU's conditionality mechanism has imposed costs on Hungary, but similar external pressure does not exist for non-EU states like Turkey, India, or El Salvador. Without external accountability, the incentive structure for autocrats is entirely domestic. |
Brookings also highlights a troubling trend in the infrastructure of democracy support: donor allocations for legislatures in Africa have declined precipitously since the late 2000s, falling from $52 million to $15 million in constant terms [8]. If democratic resilience depends on institutional capacity, and if that capacity depends in part on external investment, then the retreat of democracy-support funding is itself a driver of democratic decline. The international community has been reducing its investment in democratic institutions at precisely the moment when those institutions face their greatest threat.
Three key lessons emerge from the comparative evidence. First, democratic recovery is possible but rare — only 18 countries out of 179 are currently democratising, representing just 5 per cent of the world's population [1]. Second, recovery requires the simultaneous presence of electoral alternation, institutional allies, and civil society mobilisation — remove any one element and the process stalls. Third, timing matters: the earlier a backsliding trajectory is interrupted, the easier recovery becomes. Once institutional capture is complete, the doors to recovery close. The evidence suggests that the window for reversal is narrow and closes faster than most observers appreciate.
The Debate
Structural decline or cyclical correction?
The evidence presented in this report supports a clear conclusion: global democracy is in structural decline. But this interpretation is not universally shared ⚖ Contested. A significant body of scholarship argues that the current wave of autocratisation is a cyclical phenomenon — a correction within the long arc of democratic progress — rather than a permanent structural transformation. The debate is not merely academic. It shapes policy responses, resource allocation, and strategic priorities.
The case for structural decline rests on the weight of quantitative evidence. Twenty consecutive years of decline, across multiple independent indices, in every region of the world, affecting both new and established democracies — this is not noise in the data. The V-Dem Institute's finding that democracy for the average person has fallen back to 1978 levels [1], combined with Freedom House's documentation that only 21 per cent of the world population lives in Free countries [2], points to a fundamental realignment of the global political order. The institutional and normative foundations that sustained the post-Cold War democratic expansion — US hegemony, EU enlargement, international democracy-promotion infrastructure — have eroded or reversed.
The case for cyclical correction draws on Brookings research suggesting that "global democracy is more resilient than you may think" [8]. Proponents note that 18 nations are currently democratising, that South Korea successfully defeated an authoritarian power grab within hours, that Poland has reversed its backsliding trajectory, and that Brazil recovered from Bolsonaro-era erosion within two years. They argue that Huntington's framework of democratic "waves" — periods of expansion followed by periods of contraction — provides historical precedent for the current downturn, and that the second wave of reverse was followed by a third wave of democratic expansion that proved larger and more durable than what preceded it.
The Case for Cyclical Correction
Previous reverse waves (1922-42, 1958-75) were followed by larger democratic expansions. The third wave dwarfed both prior waves in scope and durability.
18 countries currently democratising. Poland, Brazil, Ecuador, and Lesotho have all reversed backsliding trajectories. South Korea defeated martial law within 6 hours.
Civil society mobilisation capacity has never been higher. Digital tools enable rapid coordination. South Korea's citizens were nominated for the Nobel Peace Prize for their resistance.
The EU's conditionality mechanism, IDEA's monitoring frameworks, and V-Dem's data infrastructure provide tools for early warning and response that did not exist in previous reverse waves.
Survey data consistently shows strong public support for democratic governance, even in countries experiencing backsliding. The desire for democracy has not diminished — the institutions serving it have.
The Case for Structural Decline
20 consecutive years of decline is unprecedented in the modern era. No previous reverse wave lasted this long or affected this many countries and regions simultaneously.
The US has lost its liberal democracy classification. Italy and the UK are autocratising. The institutional heartlands of post-war democratic order are eroding from within.
44 countries autocratising vs 18 democratising. 41% of world population in backsliding countries vs 5% in recovering ones. The ratio is worsening, not improving.
Modern autocrats use legal mechanisms, AI surveillance, and disinformation at industrial scale. The playbook is more sophisticated and harder to counter than any previous authoritarian model.
The US was the anchor of the post-war democratic order. Its own backsliding removes the most powerful external guarantor of democratic norms, with cascading effects on alliances, aid, and norm-setting.
The contested nature of this debate has policy implications. If the decline is cyclical, the appropriate response is patience and continued investment in democratic institutions — supporting civil society, strengthening electoral integrity, and maintaining international norms until the next wave of democratisation begins. If the decline is structural, more fundamental interventions are required: redesigning democratic institutions for the digital age, building new forms of international accountability, and addressing the underlying economic and social conditions that make populations receptive to authoritarian appeals.
The traditions underpinning America's democratic institutions are unravelling, opening up a disconcerting gap between how our political system works and long-standing expectations about how it ought to work.
— Steven Levitsky & Daniel Ziblatt, How Democracies Die, 2018One area of genuine consensus emerges: regardless of whether the trajectory is structural or cyclical, the current moment is dangerous ◈ Strong Evidence. The loss of US liberal democracy status, the reclassification of six European and North American countries as autocratising, and the collapse of press freedom to its lowest recorded level all point to a system under severe stress. Whether that stress is the prelude to a new wave of democratic expansion — as the optimists hope — or a sign of permanent structural transformation — as the data currently suggests — remains the central question in global governance.
The evidence, on balance, supports a more pessimistic reading. The asymmetry between autocratisation and democratisation — 44 countries versus 18, 41 per cent of the world's population versus 5 per cent — is not consistent with a temporary correction. The loss of the United States as a liberal democracy — the country that anchored the post-war democratic order and served as the primary external guarantor of democratic norms — is a qualitative shift, not a quantitative one. And the spread of autocratisation into Western Europe suggests that even the most deeply institutionalised democracies are vulnerable.
Yet certainty is unwarranted in either direction. Huntington's own analysis of democratic waves reminds us that apparent declines can be followed by expansions of greater magnitude. The current generation of democratic citizens — educated, connected, and politically aware — may yet prove more resilient than the institutions that currently fail them. The evidence demands vigilance, not fatalism.
What the Evidence Tells Us
The architecture of democratic resilience
Twenty years of data, four major democracy indices, and dozens of country-level case studies converge on a set of conclusions that are difficult to contest ◈ Strong Evidence. Democratic backsliding is real, accelerating, and now global in scope. It is reversible — but only under conditions that are increasingly rare. The question facing every democracy on Earth is not whether decline is possible, but whether the conditions for resilience still exist.
The first conclusion is that the autocrat's playbook is now well understood. Executive overreach, judicial capture, media consolidation, civil society suppression, and electoral manipulation — the mechanisms of democratic erosion are documented with empirical precision by V-Dem, Freedom House, International IDEA, and a generation of comparative political scientists [1] [2] [3] [7]. The playbook is consistent across continents, adaptable to local conditions, and remarkably effective. Modern autocrats do not seize power — they erode the institutions that constrain it, using legal and constitutional mechanisms that make each step individually defensible and collectively devastating.
Of 179 countries assessed by V-Dem, only 18 are currently democratising, representing just 5 per cent of the world's population [1]. Carnegie's research identifies four successful reversals (Brazil, Ecuador, Lesotho, Poland), all of which occurred before full institutional capture [5]. Once the judiciary, media, and civil society are fully co-opted — as in Hungary, Russia, and Turkey — the mechanisms of recovery are no longer functional. The evidence is clear: intervene early or face permanent erosion.
The second conclusion is that democratic resilience is not a property of institutions alone — it is a property of the relationship between institutions, civil society, and political culture. South Korea's six-hour martial law crisis demonstrates that strong institutions plus mobilised citizens plus a military that respects constitutional boundaries can defeat an authoritarian power grab in real time [9]. Hungary's trajectory demonstrates that when these elements are systematically degraded over time, no single remaining actor is strong enough to resist alone. The architecture of democratic resilience requires redundancy — multiple, independent centres of power that can check the executive even when other centres have been captured.
The third conclusion is that the loss of the United States as a liberal democracy is a qualitative transformation of the global democratic order, not merely another data point in a downward trend. The US was not just the most powerful democracy — it was the structural anchor of the post-war system of democratic norms, alliances, and institutions [1] [15]. Its own backsliding removes the most influential external guarantor of democratic standards, with cascading effects on the credibility of democracy-promotion programmes, the enforcement of international norms, and the strategic calculations of autocrats and democrats worldwide.
Democratic resilience requires three interlocking components: institutional integrity (independent judiciary, free press, functional legislature), civic capacity (organised civil society, informed citizenry, culture of democratic participation), and external accountability (international monitoring, conditionality mechanisms, peer pressure). Remove any one pillar and the structure becomes unstable. The evidence from 20 years of democratic recession suggests that all three are under simultaneous assault — and that the assault is accelerating.
The fourth conclusion is that international mechanisms for defending democracy are necessary but insufficient. The EU's conditionality regulation — freezing €15.9 billion from Hungary — represents the strongest financial tool ever deployed against democratic backsliding within an international organisation [13]. But it has not reversed Hungary's trajectory. The Article 7 procedure remains structurally paralysed by the unanimity requirement. Outside the EU, no comparable mechanism exists. The international architecture for democratic defence was designed for a world in which democratic backsliding was the exception. It is now the norm.
The fifth conclusion is that time is the critical variable. Every case study in this report confirms that the earlier a backsliding trajectory is interrupted, the more likely recovery becomes [5]. South Korea's martial law was defeated in six hours because institutions were still intact. Poland's recovery was possible because the EU exerted pressure before full institutional capture. Brazil's reversal succeeded because the democratic opposition mobilised before Bolsonaro could consolidate power. In Hungary and Turkey, where the erosion has continued for over a decade, the conditions for reversal have become exponentially more difficult to create.
Is democratic backsliding reversible? The evidence says: yes, but only if you act before the window closes. The data from V-Dem, Freedom House, International IDEA, Carnegie, and Brookings converges on a single imperative — democratic resilience is not a passive state. It requires active maintenance, vigilant civil society, independent institutions, and the political will to defend constitutional norms before they are hollowed out beyond repair.
The 20-year democratic recession has revealed a truth that the post-Cold War era obscured: democracy is not the default state of human governance. It is an achievement — fragile, contingent, and requiring constant renewal. The countries that have recovered — Poland, Brazil, South Korea — did so not because recovery was inevitable, but because specific actors made specific choices at specific moments. The countries that have not recovered — Hungary, Turkey, Tunisia — illustrate what happens when those choices are not made, or when they come too late.
The evidence demands neither optimism nor despair. It demands clarity about what is happening, precision about what works, and urgency about the narrowing window in which to act. Democracy is not dying everywhere at once. But it is retreating — faster, further, and into more unexpected places than at any point in the modern era. Whether this retreat becomes a rout or a rallying point depends on choices that are being made now, in legislatures, courtrooms, newsrooms, and streets across the world.