Authoritarianism and Democracy
The global democratic recession measured. The rise of "competitive authoritarianism." Which model is winning where. And what the actual fight looks like, beneath the slogans on every side.
(50% in 2005; the Economist Intelligence Unit puts 55% under hybrid or authoritarian regimes)
(against 18 democratizing; ten years earlier, 28 were autocratizing)
(Freedom House, 2026 report: 54 countries declined, 35 improved)
A note on framing. The democracy-versus-authoritarianism conversation is one of the most loaded in modern politics. Both the "democracies are uniquely virtuous" framing and the "authoritarianism is the future" framing carry political baggage that does not match the underlying data. A separate note on the data itself: the dominant measurement frameworks (V-Dem, Freedom House, the Economist Intelligence Unit's Democracy Index) are Western academic and policy projects. They define "democratic quality" using specific criteria (competitive elections, civil liberties, judicial independence, press freedom) that fit liberal-democratic political traditions. Non-Western governments and scholars often work from different definitions, weighting state effectiveness, social order, developmental outcomes, or popular consent in ways the Western indices do not center. China's government, for one, holds that whether a country is democratic "should be judged by its people, not dictated by a handful of outsiders." The page below uses the Western indices because they are the most widely used long-run datasets, but the measurement choice itself is not neutral. The shifts the page describes are real on the Western definition; how to weight them against other criteria is contested.
The shape of the global democratic recession
Three major academic and journalistic projects have tracked global political freedom for decades: the V-Dem Institute at the University of Gothenburg, Freedom House in Washington, and the Economist Intelligence Unit (EIU). Their methodologies differ in detail: V-Dem aggregates ratings from more than 4,000 country experts through statistical models, Freedom House uses smaller expert panels, and the EIU mixes expert assessment and public surveys. But they share common underlying data sources (national legal codes, election-monitoring reports, press-freedom rankings, civil-society litigation records) and a common conceptual framework (Robert Dahl's polyarchy criteria, with variations). They are not three fully independent measurements; their agreement is partly an artifact of applying related Western-academic frameworks to related evidence. Over two decades the direction has been consistent: Freedom House has recorded a decline in global freedom every year since 2006, and V-Dem's average country score is 10% below its 2012 peak. For 2025 they split, with the EIU finding its global average stabilizing after eight years of decline. Even four of V-Dem's five principal investigators, writing in its 2026 report, called some of its conclusions about the global decline "overstated and lacking in nuance."
Specific dimensions of the decline. Freedom of expression has been hit hardest: V-Dem counts 44 countries where it was deteriorating in 2025 and 11 where it was improving, against 7 and 25 in 2005. Independent judiciaries have been pressured or restructured in a long list of cases (Hungary before 2026, Poland before its 2023 election, India, Turkey, and on V-Dem's data the United States in 2025). Election quality was deteriorating in 22 countries and improving in 7, V-Dem finds. Tanzania's 2025 election excluded opposition candidates, Freedom House reports, and Romania's Constitutional Court annulled a 2024 presidential vote over alleged Russian interference before a 2025 rerun. Civil liberties have narrowed in many countries that still hold elections. Each of these is usually a small change in any single year and a substantial cumulative shift across a decade.
The geographic pattern. Sub-Saharan Africa has had several democratic regressions through coups (Mali, Burkina Faso, Niger and Gabon since 2020, then Madagascar and Guinea-Bissau in 2025). South and Southeast Asia have seen mixed patterns, with military rule in Myanmar, whose coup leader became president in April 2026 after a junta-run election; democratic backsliding in India under the Modi government, according to V-Dem; and continued restrictions in Cambodia, Vietnam, and Laos. The Middle East has remained mostly authoritarian with a few partial liberalizations. Eastern Europe has been mixed: Hungarian voters removed Orbán's government in April 2026 and Poland has been recovering since 2023, but V-Dem finds Slovakia, Croatia, Slovenia and Romania in decline and rates Serbia and Georgia as electoral autocracies. Latin America has been volatile but mostly remains democratic; Venezuela has had no election date since US forces seized Nicolás Maduro in January 2026. The most consequential cases - because of size and global influence - are India, Turkey, and the United States, where the democratic-quality measures show meaningful decline from previously stronger positions.
Competitive authoritarianism: the in-between regime
The classic 20th-century divide between "democracy" and "authoritarianism" has been complicated by the rise of regimes that are clearly neither. These hybrids hold elections, allow some opposition, maintain a free-ish press, and yet systematically tilt the playing field to keep the incumbent in power. The political scientists Steven Levitsky and Lucan Way called this "competitive authoritarianism" in an influential 2002 article, and the category has become increasingly useful for describing what is actually happening in many countries.
How it works in practice. Competitive authoritarian regimes hold elections that are real but not fair - opposition parties are harassed but not banned, journalists are pressured but not jailed in large numbers, courts are stacked but still occasionally rule against the government. The regime stays in power through a combination of genuine popular support, control of media, exploitation of state resources for its campaigns, and selective application of laws against opponents. The system looks democratic enough from outside that it is harder to call out than a clean dictatorship would be. Examples include Hungary under Orbán until his 2026 defeat, Turkey under Erdoğan, Russia until at least 2018 (it has hardened since), and Venezuela under Chávez and increasingly under Maduro. Whether the United States now belongs on the list is disputed (see below).
The category is large. V-Dem's nearest equivalent, electoral autocracy, is the world's most common regime type (57 countries in 2025) and covers 46% of the world's population, India included. Most democratic backsliding does not produce a coup or a clean break with democratic forms; it produces this in-between regime that is harder to fight precisely because the democratic surface remains intact.
The China model question
Throughout the post-2008 period a quiet conversation has been happening across non-Western governments about whether the "China model" - one-party rule combined with state-managed capitalism and rapid economic growth - is more attractive than Western democratic capitalism for countries trying to develop. The answer is contested.
The case for the model: China's real output per person roughly quadrupled between 2005 and 2025 and hundreds of millions of people moved out of poverty. Infrastructure was built at a pace no democracy has matched, and the state delivered some functions (large-scale construction, public security, technology development) at high effectiveness. The model appeals to governments trying to deliver development quickly and to elites who do not want to face contested elections.
The case against: China's growth has slowed from about 10% a year in the 2000s to 5.0% in 2025 (covered in the China piece on this site). The economic model that produced the 2000s boom is producing serious debt overhang and demographic stress now. The concentration of power under Xi Jinping has removed some of the safety valves earlier Chinese leaders relied on. However, the model continues to deliver in specific domains - infrastructure, public safety, technology development, and large-scale crisis management - and many of China's 1.4 billion people experience it as functional governance, not oppression. The model worked for China's specific conditions (large unified state, integration with Western export markets, technology absorption from abroad) that other countries cannot easily replicate. Poland and the Czech Republic, by contrast, grew faster as democracies than they had under communism, though whether democracy caused that success or was one of several favorable conditions is debated.
What is actually spreading. The full Chinese package - one-party rule, state-managed capitalism, integration with Chinese trade and finance - has been adopted by very few countries directly. What has spread more is specific tools: surveillance technology built in China, internet-control techniques pioneered in China, the Belt and Road infrastructure model, and most importantly the idea that developmental success does not require democratic political institutions. Whether this model is durable or whether the Chinese slowdown undermines its appeal is one of the central political-economy questions of the period.
The American backsliding question
The United States is the most consequential case in the global democracy conversation, partly because of its size and influence and partly because it has historically been a model that other democracies and aspiring democracies have looked to. The democratic-quality measurements show decline in the US over the last decade, with the contested 2020 election aftermath, the January 6, 2021 riot at the Capitol, the post-2021 challenges to election administration, and various pressures on independent institutions all contributing. In 2025 the fall steepened. V-Dem's score dropped 24%, the largest one-year fall in US history on its data. Freedom House cut the US from 84 to 81 out of 100, and the EIU's score of 7.65 out of 10 was its lowest since the index began in 2006.
Whether the United States has become a competitive authoritarian regime is disputed. Steven Levitsky, who coined the term with Lucan Way, now describes it as a form of authoritarianism. Those who reject the label point out that elections still happen, the opposition still wins (the 2024 cycle being the most recent national example), the press is still free, courts still rule against the executive, and civil society is still vigorous. The White House called V-Dem's verdict "a ridiculous claim made by an irrelevant, blatantly biased organization." The three major indices no longer agree. V-Dem's 2026 report moved the US from "liberal democracy" to "electoral democracy." Freedom House still rates the country Free. The EIU still places it among flawed democracies, the category it has occupied since 2016. The three weight executive constraints and judicial independence differently, and that is where the recent movement has been concentrated, so their disagreement is largely about which criteria carry the most weight.
What has eroded specifically. Trust in elections (covered in the Trust piece on this site) has fallen sharply, particularly among one party's supporters. Norms around peaceful transfer of power were tested in ways they had not been in modern American history. Critics say the independence of specific institutions has been removed by decision rather than by drift: in his second term Donald Trump publicly directed prosecutions of named opponents (the indictments were dismissed because the prosecutor had been unlawfully appointed), dismissed inspectors general en masse, and made about 8,000 senior civil-service posts removable at will. On V-Dem's data, Congress's constraints on the executive lost a third of their value in 2025. The administration says it is ending its predecessor's "weaponization" of law enforcement against political opponents, and each side accuses the other of using state power this way. None of these alone is a regime change; together they describe a political system under more stress than at any point in the post-WWII era.
What has held. The federal structure (50 states with substantial autonomy) has acted as a brake on centralized power. The independence of the federal courts, while tested, has continued to function in most cases - including against the president who appointed a third of the Supreme Court. That Court struck down his emergency tariffs and in December 2025 refused to let him deploy the National Guard in Illinois, though in June 2026 it also let him fire Federal Trade Commission members at will. The two-party structure has produced changes of government on schedule. Civil society organizations - business, religious, civic, professional - continue to hold the state accountable in specific ways. The deeper democratic infrastructure of the US has been more resilient than the most pessimistic readings of recent years would suggest, while remaining under more pressure than the most optimistic readings acknowledge. The event-by-event ledger of that pressure - what bent and what held under the second Trump term - is kept on the United States page.
How countries compare
Democratic quality varies dramatically across countries, and the patterns matter for understanding what is at stake. The numbers below are V-Dem's Liberal Democracy Index scores for 2025 (on a 0-1 scale), from its March 2026 report.
The takeaway: the world is more democratically diverse than the simple "democracy vs autocracy" binary suggests, with much of the action in the middle categories. The trend lines have been mostly negative for two decades; recoveries in Poland, Brazil and South Korea, and Hungary's change of government in 2026, show that reversal is possible but rare.
Why democracies erode rather than collapse
One of the major contributions of recent political science has been mapping how democratic systems actually weaken. The dominant pattern is not military coup or sudden constitutional rupture; it is gradual erosion through specific predictable steps. Understanding the steps is what makes it possible to identify backsliding while it is still happening.
Levitsky and Ziblatt's "How Democracies Die" laid out the modern playbook: weakened democracies first lose informal norms (mutual toleration of opposition, institutional forbearance) before formal rules change; then the ruling party uses available legal tools to weaken opposition (legal harassment, election-administration changes, judiciary stacking, civil-service politicization); then formal rule changes ratify the new reality. Each step is small enough to be defended individually as legitimate politics. The cumulative effect is the system becoming meaningfully different.
Specific tactics recur across cases: capturing the constitutional court or supreme court through procedural changes; weakening election administration through partisan appointments; using anti-corruption or anti-terror law against political opponents. Others include restricting press freedom through specific cases against specific outlets rather than blanket censorship, using public broadcasters as ruling-party media, and gerrymandering or election-rule changes to lock in incumbent advantage. The Hungarian playbook of 2010-2026 is the clearest single case study. Variants have appeared in Poland, Turkey, Venezuela, India, Serbia, Brazil during the Bolsonaro period, and, V-Dem argues, the United States since 2025.
What stops the erosion. The cases where backsliding has been arrested or reversed (Brazil after the 2022 election, Poland after the 2023 election, South Korea after its 2024 martial-law crisis, Hungary at the 2026 election) share several ingredients. They typically combine opposition unification across normal political differences, civil society mobilization, free institutions (independent media, courts, civil service) acting in their formal roles, international pressure (less effective than domestic factors but real), and economic stress that cuts the ruling party's support. None alone is sufficient; combinations have worked. The reversals are contested in turn: Brazil convicted Jair Bolsonaro of an attempted coup after a trial the US president called "a Witch Hunt," while President Lula said "Brazilian justice must be respected."
The paths from here
Dated September 2026, six months after V-Dem stopped counting the United States as a liberal democracy and five months after Hungarian voters removed Viktor Orbán. The democracy-versus-authoritarianism contest is not a single fight that will be won or lost in any specific year. It is a lasting condition of global politics over the next two decades. Each path below is one realistic shape that condition could take.
Continued slow erosion
The two-decade pattern of declining global democratic freedom continues. More countries drift into competitive authoritarianism. The category of "full democracies" shrinks to a smaller core in northern Europe, parts of the Anglosphere, and specific Asian and Latin American holdouts. The world becomes more autocratic on average without any single dramatic event marking the transition.
Will it happen? This is the base case if no major countervailing force emerges, and the pace would stay slow: V-Dem's principal investigators note that 175 of 179 countries saw no statistically significant change on its main index in 2025, and the EIU found scores stabilizing.
A democratic recovery wave
Following the turnarounds in Brazil (2022), Poland (2023) and South Korea (2025), Hungary's change of government (2026), and a small set of others, a broader wave of democratic recoveries reverses some of the damage of the 2010s. Specific institutional reforms (election administration protection, judicial independence reinforcement, press freedom protection, civil-service insulation) become a coordinated agenda across multiple countries. The democratic recession plateaus and partly reverses.
Will it happen? Some recovery is plausible. Hungary showed in 2026 that an entrenched government can lose, and V-Dem finds that 70% of democratization episodes that begin in autocracies succeed. But it counts only 18 democratizers, and the aid budgets that fund democracy promotion are shrinking: official development assistance fell 23% in 2025.
The US reform path
The American political system, under stress for a decade, undertakes serious institutional repair: bipartisan election-administration reform, judicial independence protections, civil-service merit reinforcement, restoration of inspector-general independence, possibly campaign-finance reform. The repair is incremental rather than transformational but stops the erosion and stabilizes the system.
Will it happen? Possible but politically difficult, and not under way. The polarization that has produced the stress also makes coordinated reform harder, and in 2025-26 policy moved the other way on inspectors general and civil-service protections. The most likely vehicle for repair is state-level action plus federal action under specific pressure conditions. A clean federal reform program seems unlikely in the immediate term.
The China model loses appeal
Chinese economic and demographic stress produces visible enough difficulties that the developmental case for one-party rule weakens. Countries that had been studying Chinese institutional models look elsewhere. The democratic-development case (Korea, Taiwan, Poland's 30-year run) reasserts itself as the more attractive model for emerging countries seeking development.
Will it happen? Possible if China continues to slow. If China's growth slows enough that it stops looking like the developmental success story it has been, the international appeal of the model fades. So far the evidence is mixed: growth has slowed to about 5% and the population is shrinking, but the IMF expects China to grow 4.6% in 2026, against 1.7% for the advanced economies.
Authoritarianism stabilizes and consolidates
The current authoritarian and competitive-authoritarian regimes hold their gains. The democratic recoveries remain isolated rather than becoming a wave. The world settles into a more multipolar system in which most people live under authoritarian or hybrid regimes, as they already do on V-Dem's count, and a minority in stable democracies. Democracy stops being the global default and becomes one of several forms of government widely accepted as legitimate.
Will it happen? One realistic outcome of the current trends, and partly under way: closed autocracies rose from 22 to 35 between 2019 and 2025, V-Dem finds. The democracy-promotion infrastructure of the post-Cold War period has weakened. The economic conditions in many democracies make institutional reform harder. The most likely shape of the 2030s and 2040s is more diverse politically than the 1990s and 2000s were.
The base case is some combination of slow continued erosion, localized recoveries, and authoritarian consolidation in many places. The specific outcomes for the largest democracies (US, India, Brazil, EU member states) will substantially shape the global trajectory; Brazil's October election and the US midterms in November are the next tests. The next decade is unusually consequential for how the democracy-versus-authoritarianism balance settles for the rest of the century.
Where serious analysts disagree
Democratic theory is one of the topics where careful researchers using similar data reach different conclusions about both the diagnosis and the prescription. Each reading below is held by named scholars worth engaging directly.
The democratic recession is real and accelerating
The patterns documented by V-Dem, Freedom House, and the EIU are not statistical artifacts or methodological problems. Real democratic erosion has been happening across many countries simultaneously for two decades, with both gradual cases (Hungary until 2026, India, Turkey, the US) and sharp ones (Russia, Belarus, Nicaragua, Venezuela, Sahel African coups). The cumulative effect is the most consequential political shift since the post-Cold War democratic wave of the 1990s, in the opposite direction.
Held by: the V-Dem Institute, the Freedom House research team, Steven Levitsky and Daniel Ziblatt (Harvard, "How Democracies Die"), and many democracy scholars. Skeptics such as Andrew Little and Anne Meng (2024) find little global decline on objective indicators, such as how often incumbents lose elections. V-Dem's researchers reply that shifting expert judgments are highly unlikely to explain the gap.
Backsliding is reversible and the recoveries are underweighted
The political-science literature has correctly identified the patterns of democratic backsliding. It has been less good at recognizing the patterns of democratic recovery. Poland in 2023, Brazil in 2022, South Korea in 2025, Hungary in 2026, and the US's institutional resilience through multiple political stress tests are all evidence that competitive authoritarianism is not a one-way ratchet. The conditions for recovery are reasonably well-understood; what is missing is sufficient attention to them.
Held by: Anna Lührmann (formerly deputy director of V-Dem, later a German minister for Europe), Larry Diamond (Stanford), and a body of comparative-politics scholarship focused on democratic resilience.
Democracies have structural economic advantages
Despite the appeal of authoritarian rapid-development models, the long-run record favors democracies. The countries that have sustained high-income status are overwhelmingly democracies. Poland and the Czech Republic grew faster as democracies than under communism, and South Korea and Taiwan, which grew fastest under authoritarian rule, kept growing after democratizing in the late 1980s. The China case is exceptional in scale but, on this view, the broader pattern is clear: institutional accountability matters for long-run prosperity.
Held by: Daron Acemoglu and James Robinson (Nobel-winning institutional economists) and the broader institutionalist tradition. Their study with Suresh Naidu and Pascual Restrepo estimates that democratization raises GDP per person by about 20% in the long run. They argue that shorter-run authoritarian successes often turn into longer-run stagnation; skeptics point to China and to the authoritarian-era growth of South Korea and Taiwan.
The democracy-vs-authoritarianism framing is too binary
The interesting political-system question of the next twenty years is not "which side wins" between democracy and authoritarianism. It is what specific institutional designs deliver what specific outcomes. The Singapore model (strict but effective and largely meritocratic single-party rule), the Nordic model (high-trust democracy with extensive welfare), the East Asian developmental-state model (democratic but technocratic), and the constitutional federalism model (US, India, Brazil, EU member states) each have different strengths. Treating "democracy" as one thing misses important variation.
Held by: Kim Lane Scheppele (Princeton), Daniel Bell (with reservations about the full Chinese model), and a body of comparative institutional analysis.
Inequality is the deeper driver
The wave of democratic backsliding tracks the wave of rising inequality (covered in the wealth-and-inequality piece on this site). Where inequality has risen sharply, support for democratic norms has weakened, populist movements have grown, and competitive authoritarian options have become attractive. Reform of democratic institutions without reform of the underlying economic conditions is treating the surface rather than the cause.
Held by: Pippa Norris (Harvard, on the cultural-economic backlash framing), Branko Milanovic (CUNY, on the global inequality patterns), and a body of economic-political scholarship.
What the evidence supports, across all these readings: the measured decline in democratic indices is real on several measurement systems, though they are not independent of one another and they disagreed about 2025. Recovery from democratic erosion is possible but requires specific institutional work. The conditions that produce democratic stress are economic as well as institutional, and the question of which political system delivers better outcomes is more productively analyzed through specific institutional design than through binary democratic-versus-authoritarian framing.
What this means for you
The quality of democratic institutions in a country shows up in everyday life through the predictability of laws, the safety of dissent, the conditions for business, and the long-run stability of the place to live and work in. A few practical observations:
If you are a citizen of a democratic country
Democratic institutions do not maintain themselves. The specific protections that the country relies on (independent courts, free press, election administration, civil-service merit, civil society organizations) require active citizens who participate, vote, contribute, and hold institutions accountable. The decline-from-the-inside path that caught Hungary, Turkey, Venezuela, and others is hard to fight from outside; it is fought primarily by domestic citizens who do unglamorous institutional work over long periods, as Hungary's 2026 election showed. Showing up matters, and a single vote weighs more in a low-turnout local or judicial election than in a high-profile national one.
If you live in a country with a hybrid or competitive authoritarian regime
The space for civil society, independent journalism, professional integrity, and quiet democratic work has not collapsed - by definition - in a competitive authoritarian regime. The work of preserving accountability under stress is real and is what makes recovery possible when political conditions allow it. The activists, judges, journalists, civil servants, and ordinary citizens who refuse to participate in the erosion are the foundation of any future recovery. The cost to them personally is real; the cumulative effect on the country's institutional health is large.
If you live in a fully authoritarian country
Citizens of authoritarian states hold a wide range of views about their own system - from active opposition to genuine support, with most people somewhere in between, focused on daily life. For those who are dissatisfied: the space for change is narrowest. Personal safety considerations are real and primary. For those who are broadly satisfied with their government's performance: the patterns described on this site are still worth understanding, because the same concentration of power that enables effective governance under a competent leader also removes accountability mechanisms that matter when leadership changes or conditions deteriorate. In both cases, the most consequential long-term action is preserving personal and institutional integrity - not collaborating beyond what is necessary, maintaining independent thought, raising children with values that outlast any particular government. Soviet-era and East European history suggests that millions of such individual choices help make societies recoverable when conditions change.
If you do business or invest internationally
Political risk has risen as a category over the last decade. Countries with deteriorating institutional quality face higher costs of capital, more capricious regulatory environments, and greater long-run uncertainty. The institutional-quality measures (V-Dem, World Justice Project Rule of Law Index, Worldwide Governance Indicators) are increasingly used as inputs to capital-allocation decisions. Decisions about where to invest, hire, and build supply chains over a 10-year horizon should now factor institutional quality more explicitly than they did in the 2000s.
If you are thinking long-term
The period from 1989 through about 2008 was unusual historically - a moment of clear democratic ascendancy. The period since has been more contested, and the period ahead will be more contested still. Operating with the assumption that democracy is the global default is increasingly out of date. Operating with the assumption that the future is necessarily authoritarian is also wrong. The realistic posture is that the contest is real and the outcome will be shaped by choices made between now and the 2040s. Engaging with that as a citizen, an investor, or simply as someone trying to understand the world is more useful than committing to either confidence or despair.
The mechanics behind this
The democracy story sits on top of three deeper mechanisms covered elsewhere on this site. If the analysis above depends on ideas you want to understand first, these fundamentals make the conversation more legible:
