Four out of five Latin Americans live under an elected government. This statistic, drawn from the UNDP’s Democracy and Development 2026 report, was reassuring ten years ago. It is troubling today. Not because the figure has declined, but because it conceals what it no longer measures: the quality of what happens within the ballot box.
An increasing number of countries on the continent are in active phases of autocratization according to the same report. Others have undergone documented democratic reversals over the past four years. Latin America is neither sliding toward authoritarianism nor consolidating its democracy. It is doing both simultaneously, in different countries, sometimes in the same country within months of each other.
The binary narrative — democracy that holds or democracy that collapses — does not account for this dynamic. What is happening is more precise, and more troubling in its precision: elected regimes that retain democratic forms while emptying them of their substance.
The Essentials
- More than 80% of Latin Americans live under elected regimes, according to the UNDP’s Democracy and Development 2026 report
- Several countries are in active phases of autocratization, including Argentina, El Salvador, Haiti, Mexico, Nicaragua, and Peru
- Brazil, Guatemala, and Bolivia have undergone documented democratic reversals since 2021
- The challenge of 2026 elections is not organizing the ballot but guaranteeing what comes after
Several Countries Are Autocratizing Without Abolishing Elections
Daniel Ortega’s Nicaragua is the clearest case. Elections continue, opposition candidates end up in prison before voting day, and official results approach 75% for the incumbent president. This is classical authoritarianism dressed up as electoral competition.
But the other cases identified by the UNDP are harder to read, and that is precisely what makes them dangerous. In El Salvador, Nayib Bukele was reelected in February 2024 with 83% of the vote, after having subordinated the legislative assembly, appointed a loyal Supreme Court, and imposed a state of exception that led to the arrest of more than 80,000 people in two years. Crime has indeed collapsed: the homicide rate has gone from one of the world’s highest to levels comparable to Europe. Polls show persistent popular approval. This is a regime that concentrates power with the active consent of its population.
Peru illustrates another mechanism. President Pedro Castillo attempted a self-coup in December 2022, was removed from office and imprisoned. His successor, Dina Boluarte, has been governing under record unpopularity and amid repeated scandals. Parliament and the executive have been mutually paralyzing each other for years. This is not authoritarianism taking hold: it is the state disintegrating. The result on civil liberties and rule of law is comparable.
Mexico under Claudia Sheinbaum inherits a constitutional reform adopted by Andrés Manuel López Obrador in 2024 that subjects federal judges to popular election. This is a structural transformation of judicial independence whose effects will be felt for a generation.
Argentina under Javier Milei commands global media attention. The ongoing economic stabilization masks growing pressure on institutions: emergency decrees used as legislative substitutes, repeated attacks on public media, frontal contestation of judicial power. As our article on the Argentine case analyzed, the record is too recent to be definitive, but the institutional signals merit being monitored separately from macro-economic results.
Brazil, Guatemala, Bolivia: What Democratic Reversals Teach
What happened in these three countries between 2021 and 2025 is less well known, and at least as instructive as the negative trends.
Brazil underwent the most dramatic test. After the January 8, 2023 assault on the state institutions’ buildings in Brasilia, the institutional response worked. The Superior Electoral Court certified Lula’s election. The armed forces did not move. State governors maintained order. Judicial proceedings against the instigators were opened. Brazil’s democratic reversal is not a miracle: it is the product of institutions solid enough to absorb a shock they had not foreseen. This detail changes everything. It was not abstract democratic norms that prevailed: it was judges, generals, and elected officials who made specific choices at specific moments.
Guatemala offers a different case, and perhaps a more replicable one. Bernardo Arévalo won the 2023 presidential election against a coalition of corrupt prosecutors and economic elites who tried to have the result annulled before he even took power. For several months, the General Prosecutor’s Office multiplied maneuvers to block the transition. What held: massive popular mobilization outside the Prosecutor’s Office, an independent press that documented every step, and coordinated international pressure, notably from the United States and the OAS. Arévalo took office in January 2024. His first term remains fragile, but he is governing.
Bolivia lived through a complete cycle: failed coup attempt in November 2019 against Evo Morales, interim government of Jeanine Áñez, 2020 elections won by Luis Arce and the MAS. Democracy survived a sequence that could have, under different circumstances, led to civil war. The fact that the 2020 elections were organized, credible, and accepted remains genuine democratic performance on a continent that has lacked recent examples of it.
These three cases share a common point that the UNDP emphasizes: the democratic reversal was not produced by institutions alone. It required a combination of citizen mobilization, active press, and international pressure. Remove one of these three elements in each case, and the result would likely be different.
The Mechanics of Democratic Hollowing
What political scientist Steven Levitsky calls competitive authoritarianism, and Anne Applebaum documents in Central Europe in Twilight of Democracy, is found in Latin America in local but recognizable forms. The illiberal kit is standardizing in Europe as well: the procedures are the same across different continents.
The sequence is now sufficiently documented to be described precisely. First, the capture of justice: appointment of loyal judges, pressure on independent prosecutors, reform of appointment methods. Then, the weakening of the press: withdrawal of licenses, judicial proceedings against journalists, acquisition of media by allies of those in power. Then control of electoral rules: party financing laws, redistricting, eligibility rules. Finally, the concentration of executive power: emergency decrees, subordination of parliament, weakening of local checks and balances.
Each of these steps is reversible taken in isolation. Together, they create a system where elections continue to be held but where the conditions of competition have been too altered for the result to still be open. The population votes. It often votes for those who have hollowed out institutions. And the figure of 80% living under elected regimes remains unchanged.
The UNDP notes in 2026 a correlation between democratic erosion and socio-economic indicators: high youth unemployment, distrust of traditional parties measured over ten years, and weak rule of law perceived in Latinobarometer polls. This is not anecdotal correlation. It indicates that democratic hollowing is not imported by malevolent leaders deceiving credulous populations. It responds to real demands: security, stability, concrete economic results. Bukele did not reduce democratic space despite his popularity. He reduced it with it.
What 2026 Will Test
Several national elections are being held in Latin America in 2026, including presidential ballots notably in Colombia (May 2026), Peru (April 2026), and Brazil (October 2026) — Chile, Bolivia, Ecuador, and Honduras having already voted in 2025. The moment is important not because results risk being contested everywhere, but because the quality of pre-electoral processes will determine the credibility of what follows.
The UNDP identifies three specific indicators to monitor in at-risk countries: the independence of electoral bodies, opposition candidates’ freedom of access to media during the campaign, and the absence of judicial pressure on candidates before the vote. These are not abstract criteria. In 2024 in Venezuela, all three were absent. The official result attributing victory to Nicolás Maduro over Edmundo González was contested by voting data released by the opposition, supported by the majority of international observers. The regime held. The election became the instrument of its own falsification.
Venezuela thus illustrates a trajectory distinct from that of countries in active autocratization phases: its regime is now classified among consolidated authoritarianisms, which says something about the direction of movement in the cases that precede it.
For Colombia notably, the stakes are clear: Gustavo Petro is nearing the end of his term with a contested record and low approval levels. The question is not the integrity of the electoral process but the capacity of institutions to produce a change of power without crisis. A Latin American democracy capable of peacefully replacing a left-wing government with a successor in a tense regional environment: that would be a useful signal.
The Actors Maintaining the Ground
The image of the continent as a space of continuous drift is inaccurate. It erases actors and institutions working precisely to maintain or rebuild the conditions of democratic play.
The Inter-American Court of Human Rights, based in San José, has rendered binding decisions against Nicaragua, Venezuela, and El Salvador on issues of arbitrary detention and freedom of the press. Its authority is contested by these three governments, all of which have threatened to withdraw from the Inter-American system, but it continues to document, condemn, and create jurisprudence usable by other courts.
Independent press organizations play a role that UNDP figures do not directly capture. In Guatemala, elPeriódico closed in 2023 under political pressure, but its teams rebuilt a newsroom in exile and continue producing coverage of the country. In Nicaragua, Confidencial operates from Costa Rica after being expelled from Managua. These journalistic resistance structures, lightweight, mobile, and often financed by international foundations, maintain independent information in contexts where national media have been captured.
The UNDP itself works in many countries in the region on strengthening electoral institutions: training of poll workers, audits of voter registries, technical support for independent commissions. This is not spectacular. This is precisely the type of invisible work that determines whether a contested election can be verified or not.
Public universities constitute another often underestimated actor. In Mexico, UNAM maintained institutional opposition to López Obrador’s judicial reforms when most political parties had capitulated. In Peru, law faculties provided the jurists who documented constitutional violations by the Castillo government. The maintenance of university autonomy in contexts of democratic erosion is not anecdotal: it is a reserve of expertise and critical legitimacy.
What Is Really at Stake by 2027
The question posed by the 2026 UNDP report is not whether Latin America will slide into authoritarianism. The data do not support that conclusion. The question is whether democratic institutions will be solid enough for the 2026 elections to produce credible changes of power in countries where they must produce them.
The Guatemalan precedent of 2023-2024 is in this respect the most instructive in the recent corpus. It shows that a legitimate president can take office against the will of a coalition of powerful actors if three conditions are met: citizen mobilization capable of sustaining pressure over time, a coherent international community in its support, and internal judicial institutions where at least one chamber or court remains uncaptured. These three conditions are not given in advance. They are built, or they are allowed to degrade.
The real test of Latin American democracies in 2026 is not election day. It is the day after, and the three months that follow.
Sources
- UNDP, Democracy and Development — Report 2026, https://www.undp.org/latin-america/democracy-and-development-report-2026
- Latinobarometer, Report 2023, https://www.latinobarometro.org
- V-Dem Institute, Democracy Report 2025, https://v-dem.net/democracy_reports.html
- Organization of American States (OAS), electoral observation reports 2023-2024, https://www.oas.org
- Steven Levitsky and Lucan Way, Competitive Authoritarianism: Hybrid Regimes After the Cold War, Cambridge University Press, 2010
- Anne Applebaum, Twilight of Democracy: The Seductive Lure of Authoritarianism, Doubleday, 2021
- UNDP, Democracy and Development Report 2026 — Official page, https://www.undp.org/latin-america/publications/democracy-and-development-report-democracies-under-pressure-reimagining-future-democracy-latin-america-and-caribbean-2026
- TSE El Salvador — official Bukele 2024 result, https://www.tripfoumi.com/blog/2024/02/10/salvador-officialisation-de-la-large-reelection-du-president-bukele/
- Congressional Research Service — State of Exception El Salvador, https://www.congress.gov/crs_external_products/IN/PDF/IN12510/IN12510.2.pdf
- Wikipedia — Peru 2022 Self-Coup Attempt, https://en.wikipedia.org/wiki/2022_Peruvian_self-coup_attempt
- Le Devoir / France 24 — Mexico Judicial Reform 2024, https://www.ledevoir.com/monde/ameriques/819623/senat-mexique-adopte-reforme-judiciaire-controversee-malgre-forte-hostilite
- Wikipedia — January 8, 2023 Brazil, https://en.wikipedia.org/wiki/2023_Brazilian_coup_d%27%C3%A9tat_attempt
- Wikipedia — 2023 Guatemalan Presidential Election, https://en.wikipedia.org/wiki/2023_Guatemalan_general_election
- Wikipedia — 2020 Bolivian General Election, https://en.wikipedia.org/wiki/2020_Bolivian_general_election
- U.S. Department of State — Venezuela 2024 Election, https://www.state.gov/the-venezuelan-electoral-process/
- Espaces Latinos / Jean-Jaurès — Latin American Elections 2026, https://www.espaces-latinos.org/archives/131050