Pedro Graterol
ENESPT
Latinobarómetro / 1995–2024The story and the research

Who Do Latin Americans Trust?

Nearly 30 years of survey data reveal a more complicated picture of political trust in Latin America: confidence in one institution can rise while confidence in another falls.

Pedro Graterol

Start reading ↓

Latin America’s crisis of trust is not one story.

Latin America is often described as facing a crisis of political trust. Citizens are losing confidence in their institutions, the argument goes, raising concerns about the future of democracy across the region.

But what happens when we stop treating political trust as a single number?

An analysis of nearly three decades of Latinobarómetro surveys across 18 countries reveals a more complicated picture: confidence in different institutions can move in different directions, even within the same country. Congress can lose trust while the police gain it. Political parties can recover while the military declines. And what looks like a recovery over a few years can still be a longer-term decline.

This does not mean the crisis of political trust is imaginary. It means there is no uniform story of institutional confidence across Latin America. To understand what is happening, we need to ask a more precise question: who do Latin Americans trust, and how has that changed over time?

What does it actually mean to trust a political institution?

Imagine someone who disapproves of the government, has little confidence in Congress and still firmly prefers democracy to any alternative. Those answers are compatible. One evaluates the people currently governing; another evaluates an institution; the third concerns the kind of political system the person wants. A survey about political trust becomes more revealing when we keep those questions apart.

David Easton’s 1965 account of political support offers a useful starting point. Specific support responds to what political authorities do and what people receive or expect from them. Diffuse support refers to a more enduring attachment that can help a political system withstand dissatisfaction with particular decisions or leaders. Someone can dislike a government’s performance without withdrawing all support from the wider political order. This is a theoretical distinction, not a one-to-one translation of our survey variables: confidence in Congress or the police does not automatically measure Easton’s diffuse support.

Pippa Norris’s work in 1999 and 2011 separates the objects of political support further. People can evaluate their political community, democratic principles, the performance of the regime, its institutions and its political actors differently. Think of a citizen who values belonging to a country, supports democratic elections, feels disappointed with how democracy works and distrusts a party. Calling that person simply “trusting” or “distrustful” would conceal most of the story.

The four concepts shown here address different parts of that picture. Government approval concerns the administration in office. Institutional confidence asks about a particular institution. Satisfaction with democracy concerns how democracy works. Support for democracy concerns its desirability as a political system. These attitudes can be connected, but a change in one does not supply the answer to the others.

Russell Dalton’s work on democratic citizenship and political support adds another possibility: strong democratic expectations can make citizens more demanding of the institutions meant to fulfill them. Criticism may reflect disappointment with representation, responsiveness or performance rather than opposition to democracy itself. That possibility is not a diagnosis of every dissatisfied respondent. Low confidence alone cannot distinguish a demanding democrat from someone who rejects democratic government.

These distinctions lead directly to this investigation’s design. Congress and parties represent citizens and organize political competition. Courts, police and armed forces have different responsibilities and different relationships with the public. Citizens may bring different expectations and experiences to each judgment. We therefore follow five institutions separately, then examine how their confidence levels relate to government approval. Easton, Norris and Dalton help frame the questions; the surveys show the patterns. Explaining the causes of those patterns requires evidence beyond the trajectories themselves.

Methods and qualifications ↗
457,499interviews
18countries
24observed waves

Archive-wide totals. Availability varies by institution, country and analysis.

Start with a contradiction

Flag of BrazilBrazil

2010–2024 / 2018–2024

Brazil’s congressional recovery depends on when you start counting.

Brazil’s confidence in Congress fell 18.2 percentage points between 2010 and 2024. Looking only from 2018 to 2024 gives an increase of 14.6 points. Recovery is visible; so is the distance from the earlier level.

The two windows share the same destination: 2024. What changes is the starting point. The bars show Congress below its 2010 level but above its 2018 level. A recovery can therefore be real without restoring the confidence seen earlier.

The historical line adds what the endpoints leave out. Move the selected wave along the chart to follow the rise and fall between surveys. The early 1995 observation comes from a partially urban sample, so it provides historical perspective with a different coverage boundary.

This investigation follows Congress, political parties, the judiciary, police and armed forces separately. Positive confidence means reporting much or some confidence, among people giving a substantive answer. A person who does not know or does not answer is not counted as expressing distrust.

Methods and qualifications ↗
  • Congress
  • Political parties
  • Judiciary
  • Police
  • Armed forces

Brazil · 2010–2024

Change · pp

Congress-18.2Political parties-8.9Judiciary-16.4Police+4.0Armed forces-18.10
Change in positive-confidence shares, using exact observed endpoints. Institution-specific denominators. Direction is not a significance test. Source: COUNTRY_INSTITUTION_PERIODS.csv.

The record starts before 2010

The first observation is not always the same starting line

The archive spans 1995–2024: 457,499 interviews, 18 countries and 24 observed survey waves. The earlier years matter because a comparison beginning in 2010 can conceal the climb that preceded a subsequent fall. But a long record is not automatically a fully comparable national history.

Brazil makes the distinction tangible. Positive confidence in Congress was 27.2% in the covered 1995 sample and 27.3% in 2024. Those similar endpoints must not be read as evidence that national confidence stayed unchanged. The early sample had partial urban coverage, and the intervening observations moved substantially.

The calendar shows what was actually observed. Countries enter at different points; Nicaragua ends in 2020. Latinobarómetro interviews different people in successive waves. Missing years remain empty, and the armed-forces histories of Costa Rica and Panama are not presented as comparable military series.

Historical depth therefore adds two things: more of the trajectory, and a clearer view of the limits of the comparison. In the explorer, the full available record remains accessible alongside shorter windows; its early observations retain their coverage warnings.

Methods and qualifications ↗

The survey calendar

19952024
Argentina
Bolivia
Brazil
Chile
Colombia
Costa Rica
Dominican Republic
Ecuador
El Salvador
Guatemala
Honduras
Mexico
Nicaragua
Panama
Paraguay
Peru
Uruguay
Venezuela
One dot per observed country-wave. These are nominal survey years, not verified fieldwork dates. Source: COUNTRY_TRAJECTORIES.csv.

Five institutions, one country

Flag of MexicoMexico

2010–2024

Four institutions gained confidence. The military lost it.

Between 2010 and 2024, Mexico recorded increases in confidence in Congress, parties, the judiciary and police, alongside a decline in the armed forces. Parties gained 11.0 percentage points; police gained 7.6. Even the institutions often grouped as representative or security institutions did not behave as interchangeable objects.

The country chart makes the split visible: four bars move to the positive side of zero, while the armed-forces bar moves to the negative side. The biggest gain shown here belongs to parties, not to a security institution.

That is why a single national score would be an incomplete description. The question is not only whether confidence rose, but which parts of the state gained it. These movements do not identify what caused the change; they establish the institutional pattern an explanation would need to address.

Mexico · 2010–2024

Change · pp

-200+20Congress+2.8Political parties+11.0Judiciary+7.2Police+7.6Armed forces-4.4
Change between observed endpoints, in percentage points; each institution uses its own substantive-response denominator. Latinobarómetro. Methods and qualifications ↗

Brazil provides a contrast. Over the same years, Congress, parties, the judiciary and armed forces all lost confidence while police gained. The heatmap shows the components: it is the combination of movements, not one institution chosen in isolation, that describes each country.

Other countries moved more broadly. Bolivia, Colombia, Honduras and Venezuela recorded negative point changes across all five institutions; the Dominican Republic and El Salvador recorded positive ones. These are descriptions of 2010–2024, not permanent country identities or claims that every change is statistically distinguishable from zero.

Police improved relative to Congress in all 17 countries with matched endpoints. Yet that common relative result includes countries where both institutions declined. Police versus armed forces, meanwhile, split eight countries in one direction and seven in the other. The record does not describe a uniform transfer of confidence from political institutions to security institutions.

Flag of ColombiaColombia

2010–2024

Police lost confidence, even as they gained ground on Congress.

Colombia adds a different test. From 2010 to 2024, all five institutions recorded declines in confidence. A relative advantage for police therefore does not necessarily describe a security institution gaining public confidence.

Congress lost 16.2 percentage points, compared with 2.9 for police. Both bars point toward declining confidence, but their lengths are very different. Police can improve their position relative to Congress while becoming less trusted in absolute terms.

This is a useful test of the regional story. A widening institutional gap does not necessarily mean that citizens are rallying behind the institution that finishes ahead. Sometimes it means that another institution has fallen much further.

Colombia · 2010–2024

Change · pp

-200+20Congress-16.2Political parties-13.6Judiciary-13.4Police-2.9Armed forces-7.0
Change between observed endpoints, in percentage points; each institution uses its own substantive-response denominator. Latinobarómetro. Methods and qualifications ↗
Flag of UruguayUruguay

2018–2024

Congress recovered. Military confidence fell.

Uruguay challenges the opposite shorthand: between 2018 and 2024, Congress gained 15.7 points, parties 14.5 and the judiciary 11.0, while police changed little and armed forces fell 5.3. Representative recovery and security recovery are different claims.

The strongest gains in this window belong to Congress and political parties, with the judiciary also rising. Police confidence is nearly unchanged, while confidence in the armed forces falls. The chart gives the institutions separate space instead of assigning Uruguay one direction.

Uruguay consequently complicates a simple contrast between political institutions in retreat and security institutions on the rise. Its recent pattern puts representative recovery at the center. That description belongs to 2018–2024; it is not a permanent label for the country.

Uruguay · 2018–2024

Change · pp

-200+20Congress+15.7Political parties+14.5Judiciary+11.0Police-0.0Armed forces-5.3
Change between observed endpoints, in percentage points; each institution uses its own substantive-response denominator. Latinobarómetro. Methods and qualifications ↗
Methods and qualifications ↗
  • Congress
  • Political parties
  • Judiciary
  • Police
  • Armed forces

Change · 2010–2024

Compare countries · percentage points
CountryCongressPolitical partiesJudiciaryPoliceArmed forces
Argentina-14.5-3.1-12.4+7.1+16.3
Bolivia-9.7-6.9-12.3-7.0-16.5
Brazil-18.2-8.9-16.4+4.0-18.1
Chile-22.3-10.1-19.3+1.1+2.0
Colombia-16.2-13.6-13.4-2.9-7.0
Costa Rica-23.2-9.1-1.7+15.1—
Dominican Republic+10.5+5.6+13.4+23.2+24.1
Ecuador-6.0-8.1-4.0-0.7+12.5
El Salvador+16.4+5.1+30.0+34.5+11.3
Guatemala+2.8-0.7+5.0+9.6+2.2
Honduras-25.7-11.5-15.1-9.1-4.7
Mexico+2.8+11.0+7.2+7.6-4.4
Nicaragua—————
Panama-16.1-16.9-1.6+4.7—
Paraguay-5.9-9.2+0.9+7.2-3.1
Peru-6.9-4.5+3.6+10.9+13.7
Uruguay-13.7-9.5-8.8+5.6+15.8
Venezuela-30.9-23.4-19.9-8.8-23.9
Change in positive-confidence shares, using exact observed endpoints. Institution-specific denominators. Direction is not a significance test. —: Unavailable. Source: Latinobarómetro / COUNTRY_INSTITUTION_PERIODS.csv.

Brazil, reconsidered

The recovery is one stretch of a longer history

The longer record helps explain the arithmetic. Congress gained confidence between 2005 and 2010, then fell sharply by 2018. The recent rise starts after that fall. An endpoint comparison can summarize one part of this sequence while hiding another.

Parties also recover in 2018–2024, gaining 9.5 points. Police falls 2.9 points and armed forces fall 11.8. The institutions do not even share the same recent direction within Brazil.

Methods and qualifications ↗
Flag of BrazilBrazil

Confidence levels

Full available history · Observed years: 1995–2024

0204060801001995200020052010201520202024
Explore observed survey waves2018
19952024
Congress12.7 %
Political parties6.3 %
Judiciary33.7 %
Police47.6 %
Armed forces59.6 %
Samples, uncertainty and sources

Source-weighted share among substantive responses. Missing waves are not zero. Dots are observed surveys, not annual estimates.

Congress · Valid responses: 1,165

Political parties · Valid responses: 1,186 · Additional institutional-referent and comparability cautions apply.

Judiciary · Valid responses: 1,171 · Additional institutional-referent and comparability cautions apply.

Police · Valid responses: 1,189

Armed forces · Valid responses: 1,170 · Additional institutional-referent and comparability cautions apply.

Source-weighted share among substantive responses. Missing waves are not zero. Source: Latinobarómetro · COUNTRY_TRAJECTORIES.csv. Methods and qualifications ↗
Values and sample sizes
Brazil · Confidence levels · 1995–2024
Survey waveInstitutionsEstimateValid responses
1995Congress27.2 %1148
1995Political parties17.4 %1170
1995Judiciary40.8 %1141
1995Police34.0 %1179
1995Armed forces60.8 %1157
1996Congress20.0 %1060
1996Political parties16.3 %1071
1996Judiciary41.4 %1068
1996Police26.6 %1076
1996Armed forces63.6 %1074
1997Congress28.1 %975
1997Political parties18.4 %979
1997Judiciary45.1 %972
1997Police32.0 %993
1997Armed forces60.2 %981
1998Congress23.4 %977
1998Political parties19.8 %989
1998Judiciary41.8 %989
1998Police24.2 %993
1998Armed forces52.5 %982
2000Congress19.0 %963
2000Political parties12.6 %981
2000Judiciary36.4 %973
2000Police29.0 %989
2000Armed forces59.7 %977
2001Congress24.7 %939
2001Political parties20.7 %968
2001Judiciary40.2 %963
2001Police32.2 %981
2001Armed forces54.9 %966
2002Congress26.2 %947
2002Political parties12.8 %969
2002Judiciary33.7 %964
2002Police37.8 %984
2002Armed forces63.1 %963
2003Congress29.4 %1157
2003Political parties15.7 %1188
2003Judiciary44.2 %1165
2003Police36.9 %1195
2003Armed forces56.4 %1158
2004Congress36.5 %1149
2004Political parties22.8 %1189
2004Judiciary49.7 %1160
2004Police36.3 %1194
2004Armed forces63.0 %1160
2005Congress27.5 %1152
2005Political parties15.8 %1193
2005Judiciary42.7 %1153
2005Police39.0 %1189
2005Armed forces62.6 %1167
2006Congress32.7 %1135
2006Political parties22.7 %1174
2006Judiciary46.1 %1164
2006Police42.5 %1186
2006Armed forces69.0 %1167
2007Congress24.4 %1158
2007Political parties16.1 %1191
2007Judiciary42.0 %1164
2007Police35.8 %1197
2007Armed forces68.5 %1181
2008Congress34.9 %1162
2008Political parties20.7 %1180
2008Judiciary42.8 %1163
2008Police39.6 %1197
2008Armed forces62.6 %1175
2009Congress32.8 %1161
2009Political parties20.8 %1184
2009Judiciary43.0 %1164
2009Police34.4 %1196
2009Armed forces62.8 %1176
2010Congress45.5 %1156
2010Political parties24.7 %1169
2010Judiciary53.0 %1161
2010Police40.7 %1186
2010Armed forces65.9 %1157
2011Congress31.3 %1172
2011Political parties15.9 %1197
2011Judiciary39.8 %1181
2011Police34.8 %1201
2011Armed forces63.6 %1192
2013Congress29.2 %1166
2013Political parties21.6 %1168
2013Judiciary42.9 %1156
2013Police44.2 %1192
2013Armed forces68.9 %1172
2015Congress19.4 %1193
2015Political parties10.4 %1221
2015Judiciary33.2 %1194
2015Police37.1 %1236
2015Armed forces56.7 %1184
2016Congress13.4 %1160
2016Political parties5.3 %1189
2016Judiciary39.9 %1163
2016Police39.5 %1194
2016Armed forces57.4 %1155
2017Congress11.0 %1163
2017Political parties6.7 %1186
2017Judiciary27.8 %1163
2017Police34.5 %1190
2017Armed forces52.4 %1148
2018Congress12.7 %1165
2018Political parties6.3 %1186
2018Judiciary33.7 %1171
2018Police47.6 %1189
2018Armed forces59.6 %1170
2020Congress23.9 %1167
2020Political parties13.1 %1186
2020Judiciary37.0 %1170
2020Police48.2 %1194
2020Armed forces58.0 %1176
2023Congress30.8 %1160
2023Political parties18.6 %1185
2023Judiciary37.5 %1170
2023Police48.2 %1182
2023Armed forces49.7 %1177
2024Congress27.3 %1149
2024Political parties15.9 %1185
2024Judiciary36.6 %1160
2024Police44.7 %1195
2024Armed forces47.8 %1180

The government is not every institution

Flag of ArgentinaArgentina

2024

Government approval reaches beyond elected institutions.

Do people who approve of their government also tend to trust the country’s institutions more?

The survey asks two different things: whether people approve or disapprove of the government, and whether they have confidence in institutions such as Congress or the armed forces. Here, we compare the answers of government approvers and disapprovers.

In Argentina in 2024, people who approved of the government were more likely to express confidence in Congress. After accounting for age and recorded sex, their estimated probability of expressing confidence was 15.5 percentage points higher than that of people who disapproved.

For the armed forces, the difference was 33.6 percentage points—more than twice as large. These numbers describe the distance between the two groups. They do not mean that 33.6% of Argentines trusted the military, or that confidence rose by that amount.

The contrast extends beyond elected institutions: the estimated difference was larger for the armed forces than for Congress. All five institutions examined in this Argentine survey had positive differences, but their size varied.

The historical models compare government approvers and disapprovers separately within each country and survey wave. Precise national extension summaries remain incompletely documented and are not used here.

These are comparisons between groups of respondents, not an experiment or evidence that approving a government causes confidence. How people evaluate their government and its institutions is connected, but the answers are not interchangeable.

A percentage-point difference is the distance between two estimated probabilities. The comparison here is between government approvers and disapprovers interviewed in the same survey, accounting for age and recorded sex. It is not a comparison of Argentina before and after a change of government.

The larger military gap gives the case its significance. Evaluations of the incumbent reach beyond elected institutions in these responses. The association could reflect shared evaluations or other political relationships; the model does not isolate a causal mechanism.

Methods and qualifications ↗

Argentina, 2024

Adjusted difference (percentage points)

Congress
+15.5pp
Armed forces
+33.6pp
Difference in the estimated probability of expressing confidence: government approvers minus disapprovers, accounting for age and recorded sex. Percentage points, not overall confidence levels.
See all five institutions and uncertainty

Common sample: 1,091 respondents. National weights; age/recorded-sex adjustment. Intervals use the inherited conditional uncertainty estimates.

Argentina, 2024 · Adjusted difference (percentage points)
InstitutionsAdjusted difference (percentage points)95% interval
Congress+15.59.8 / 21.2
Political parties+10.14.9 / 15.3
Judiciary+21.516.0 / 27.0
Police+28.022.0 / 34.1
Armed forces+33.627.8 / 39.4
Explore approval associations

The historical models compare government approvers and disapprovers separately within each country and survey wave. Precise national extension summaries remain incompletely documented and are not used here.

Sources: inherited approval estimates; frozen research report, p. 13. Methods and qualifications ↗

Behind the comparisons

Behind these charts is a much larger research effort.

The analytical archive contains 457,499 respondent records from 24 observed survey waves in 18 countries, organized into 413 country–wave samples spanning 1995–2024. These are archive-wide counts, not the sample for every chart or model. The work behind the publication was to turn that uneven historical record into comparisons whose definitions and boundaries can be inspected.

A shared column name is only a starting point. The project assembled source files and metadata, identified candidate variables, checked wording and response scales, established harmonization and exclusion rules, and constructed eligible series. Consistency and robustness checks then tested how sensitive the findings were to documented analytical choices. The decision trail matters as much as the volume of records.

The contribution has two parts. Empirically, the project makes institutional trajectories easier to compare across countries and periods. Methodologically, it provides a documented framework for deciding which historical comparisons are defensible. Public code, methods and the audit make that framework inspectable; some reproduction still depends on external data and incomplete historical artifacts.

The archive behind the comparisons

457,499
interview records in the frozen harmonized archive
18
countries
24
observed survey waves
413
country-wave samples
5
institutional confidence measures

1995–2024 describes the archive, not complete coverage for every measure. The interview count is not a separate total of unique people followed over time. Valid answers and model samples are smaller and vary by question. No separate source-versus-harmonized respondent count is asserted here.

  1. Identify

    Locate source questions, labels and codes for each institution and wave. Preserve the original item and its provenance; a matching variable name is only a candidate link.

  2. Compare

    Check wording, response scales and national institutional referents against the documentation. Retain comparability classifications and measurement segments; availability alone does not establish equivalence.

  3. Standardize

    For confidence, combine “much” and “some” (categories 1/2) as positive, against “little” and “none” (3/4). Keep non-substantive responses distinct where the source permits.

  4. Validate

    Establish country–wave–institution coverage and eligibility. An unavailable item stays missing; incompatible episodes remain separate or excluded. Costa Rica and Panama, for example, are not assigned zero military confidence when a comparable armed-forces referent is absent.

  5. Document

    Freeze the harmonized dataset and retain coding, clearance, sample restrictions and source links. Use original national weights within each country-wave and substantive answers as the primary denominator; national weights are not regional population weights.

Why does a study beginning in 1995 sometimes start its comparisons in 2010?

The archive begins in 1995, but a comparison needs more than an early observation. It needs the relevant institutions and countries observed at both chosen endpoints, with their measurement restrictions retained. The documented 2010–2024 comparison pairs 17 countries for Congress, parties, judiciary and police, and 15 for armed forces. It is one analytical window, alongside earlier histories and other endpoint comparisons; 2010 is not a universal threshold of survey quality. Brazil illustrates the distinction: its 1995 Congress observation is available, but the early sample had partial urban coverage. Comparing it with 2024 therefore describes the covered samples, not an unchanged national population. Earlier observations have not been discarded. They remain in the historical explorer where defensible, with missing waves and design boundaries preserved. A common comparison window makes the country set explicit; the full record shows what that window leaves outside.

Harmonization makes definitions explicit; it does not make three decades of surveys identical. These are different respondents at each wave. Sampling frames, territorial coverage and unavailable survey-design details still limit inference. The findings describe confidence and associations, not individual change or causal effects.

Methods and qualifications ↗

One definition, carefully checked

1 + 2

Much or some confidence

1
3 + 4

Little or no confidence

0
—

Don’t know / no response

Kept outside the substantive denominator
Conceptual illustration of the research definition, not a new survey observation. Unavailable items remain unavailable.

What confidence does not tell us

Distrust of an institution is not a vote against democracy

Being satisfied with how democracy works is different from supporting democracy as a political principle. Confidence in Congress is different again. Treating these as one attitude would erase a distinction the evidence preserves.

Flag of ChileChile

2018–2024

Police gained confidence. Courts and parties lost it.

Chile shows why those distinctions matter. Between 2018 and 2024, police confidence rose 14.7 points and military confidence 6.4, while judiciary confidence fell 8.6 and party confidence 2.4. This institutional configuration does not establish whether Chileans became more committed to democracy.

The change chart shows police and armed forces moving upward while courts and parties move downward. Those contrasting directions describe how confidence was distributed across institutions over this period.

To ask whether Chileans became more satisfied with democracy, or more committed to democratic principles, we would need to examine those separate questions. Chile’s institutional pattern makes that distinction concrete: confidence in one part of the state cannot stand in for a judgment about the whole democratic system.

Chile · 2018–2024

Change · pp

-200+20Congress+0.9Political parties-2.4Judiciary-8.6Police+14.7Armed forces+6.4
Change between observed endpoints, in percentage points; each institution uses its own substantive-response denominator. Latinobarómetro. Methods and qualifications ↗

A person can disapprove of the government, distrust Congress and still prefer democracy to any alternative. These judgments may overlap, but they ask different questions. Government approval evaluates the administration in office. Institutional confidence concerns a particular institution. Satisfaction with democracy evaluates how democracy works; support for democracy concerns its desirability as a political system.

The distinction becomes more consequential when the question shifts from confidence to political legitimacy.

Flag of VenezuelaVenezuela

2010–2024 / 2018–2024

Confidence figures cannot settle the question of legitimacy.

Venezuela also resists a single description. All five confidence measures declined between 2010 and 2024. Yet from 2018 to 2024, police rose 10.2 points, armed forces 7.2 and parties 6.3, while Congress fell 0.9. These reported responses cannot establish democratic legitimacy, nor can the surveys alone establish that respondents concealed their preferences. Political response environments and incomplete national-questionnaire verification constrain stronger interpretations. The seven countries used in this presentation illustrate distinct puzzles; the explorer retains all 18.

The line chart brings the different windows into the same view. The rise after 2018 in several institutions sits inside a longer record in which all five ended below their 2010 levels. As in Brazil, the starting year changes which part of the history becomes visible.

There is also an interpretive boundary. Reported confidence in an institution and democratic legitimacy are different judgments. These surveys record the answers people gave; they do not reveal hidden preferences or establish how the political response environment shaped each answer. Venezuela sharpens the question that runs through the presentation: confidence in what, compared with when, and with what meaning?

Flag of VenezuelaVenezuela

Confidence levels

Selected window: 2010–2024 · Observed years: 2010–2024

0204060801002010201520202024
Explore observed survey waves2024
20102024
Congress19.7 %
Political parties20.5 %
Judiciary19.5 %
Police22.3 %
Armed forces26.2 %
Samples, uncertainty and sources

Source-weighted share among substantive responses. Missing waves are not zero. Dots are observed surveys, not annual estimates.

Congress · Valid responses: 1,174

Political parties · Valid responses: 1,163 · Additional institutional-referent and comparability cautions apply.

Judiciary · Valid responses: 1,167 · Additional institutional-referent and comparability cautions apply.

Police · Valid responses: 1,182

Armed forces · Valid responses: 1,182 · Additional institutional-referent and comparability cautions apply.

Source-weighted share among substantive responses. Missing waves are not zero. Source: Latinobarómetro · COUNTRY_TRAJECTORIES.csv. Methods and qualifications ↗
Values and sample sizes
Venezuela · Confidence levels · 2010–2024
Survey waveInstitutionsEstimateValid responses
2010Congress50.6 %1171
2010Political parties43.9 %1170
2010Judiciary39.4 %1157
2010Police31.1 %1188
2010Armed forces50.1 %1167
2011Congress43.5 %1176
2011Political parties35.5 %1183
2011Judiciary37.6 %1179
2011Police26.4 %1194
2011Armed forces46.1 %1180
2013Congress41.9 %1168
2013Political parties43.0 %1173
2013Judiciary40.1 %1164
2013Police39.7 %1177
2013Armed forces54.9 %1186
2015Congress29.1 %1192
2015Political parties21.6 %1198
2015Judiciary25.9 %1186
2015Police16.1 %1190
2015Armed forces36.4 %1196
2016Congress40.3 %1190
2016Political parties20.5 %1187
2016Judiciary19.0 %1188
2016Police15.8 %1199
2016Armed forces28.0 %1199
2017Congress37.7 %1174
2017Political parties30.7 %1183
2017Judiciary26.6 %1180
2017Police28.6 %1192
2017Armed forces34.8 %1194
2018Congress20.6 %1178
2018Political parties14.2 %1182
2018Judiciary17.9 %1175
2018Police12.1 %1195
2018Armed forces19.1 %1191
2020Congress19.7 %1168
2020Political parties15.1 %1178
2020Judiciary18.3 %1170
2020Police13.6 %1187
2020Armed forces20.6 %1188
2023Congress23.6 %1179
2023Political parties16.3 %1183
2023Judiciary24.6 %1180
2023Police21.8 %1196
2023Armed forces26.5 %1192
2024Congress19.7 %1174
2024Political parties20.5 %1163
2024Judiciary19.5 %1167
2024Police22.3 %1182
2024Armed forces26.2 %1182
Methods and qualifications ↗

Different democratic evaluations, different relationships

Four questions, four objects of evaluation

Institutional confidence

Confidence in a particular institution.

Government approval

Evaluation of the government in office.

Satisfaction with democracy

Evaluation of how democracy works.

Support for democracy

Preference for democracy as a political system.

Related judgments. No single answer substitutes for the others. Methods and qualifications ↗

What these comparisons can tell us

These surveys interview different people in successive waves. A national increase describes a change in the distribution of answers, not a change followed within the same individuals. Missing surveys stay missing; line connections are guides between observations, not measurements for every intervening year.

Questions, response categories, institutional names and sample coverage can change. Harmonization makes those differences explicit and excludes incompatible episodes; it cannot make every country and year perfectly equivalent. The early Brazilian sample and Argentina’s 2020 survey-mode break illustrate why the qualifications travel with the data.

Reported confidence is an evaluation of a specific institution. Survey response biases and political environments may affect answers, but these data alone do not measure those effects or reveal respondents’ sincerity. Confidence cannot automatically be treated as democratic legitimacy. Approval models describe adjusted associations, not causal effects, and their uncertainty does not capture every feature of the original survey designs.

The audit reproduced the core confidence estimates and approval coefficients. Some later summaries lack complete analytical recipes and have been removed from the numerical story. The researchers’ section preserves those verification boundaries, the full measurement limitations and the external-data requirements for reproduction.

Methods and qualifications ↗

Brazil’s recovery is also part of a longer decline. Mexico’s gains leave the military behind. In Argentina, the difference between government approvers and disapprovers extends beyond elected institutions. Each finding changes the meaning of a simple claim that a country is gaining—or losing—trust.

Institutional confidence matters to the story of democracy, but it is not a substitute for it. Satisfaction with democracy, support for democratic principles and confidence in particular institutions can be related without telling the same story.

A country does not have one trust story. It has institutions whose fortunes can separate, histories whose starting points matter, and judgments about democracy that cannot be read from a single confidence score.

08 / Explore the data

Who gets confidence?

Historical survey coverage and institutional referents vary. Consult the methods.

Institutions
Flag of BrazilBrazil

Confidence levels

Full available history · Observed years: 1995–2024

0204060801001995200020052010201520202024
Explore observed survey waves2018
19952024
Congress12.7 %
Political parties6.3 %
Judiciary33.7 %
Police47.6 %
Armed forces59.6 %
Samples, uncertainty and sources

Source-weighted share among substantive responses. Missing waves are not zero. Dots are observed surveys, not annual estimates.

Congress · Valid responses: 1,165

Political parties · Valid responses: 1,186 · Additional institutional-referent and comparability cautions apply.

Judiciary · Valid responses: 1,171 · Additional institutional-referent and comparability cautions apply.

Police · Valid responses: 1,189

Armed forces · Valid responses: 1,170 · Additional institutional-referent and comparability cautions apply.

Source-weighted share among substantive responses. Missing waves are not zero. Source: Latinobarómetro · COUNTRY_TRAJECTORIES.csv. Methods and qualifications ↗
Values and sample sizes
Brazil · Confidence levels · 1995–2024
Survey waveInstitutionsEstimateValid responses
1995Congress27.2 %1148
1995Political parties17.4 %1170
1995Judiciary40.8 %1141
1995Police34.0 %1179
1995Armed forces60.8 %1157
1996Congress20.0 %1060
1996Political parties16.3 %1071
1996Judiciary41.4 %1068
1996Police26.6 %1076
1996Armed forces63.6 %1074
1997Congress28.1 %975
1997Political parties18.4 %979
1997Judiciary45.1 %972
1997Police32.0 %993
1997Armed forces60.2 %981
1998Congress23.4 %977
1998Political parties19.8 %989
1998Judiciary41.8 %989
1998Police24.2 %993
1998Armed forces52.5 %982
2000Congress19.0 %963
2000Political parties12.6 %981
2000Judiciary36.4 %973
2000Police29.0 %989
2000Armed forces59.7 %977
2001Congress24.7 %939
2001Political parties20.7 %968
2001Judiciary40.2 %963
2001Police32.2 %981
2001Armed forces54.9 %966
2002Congress26.2 %947
2002Political parties12.8 %969
2002Judiciary33.7 %964
2002Police37.8 %984
2002Armed forces63.1 %963
2003Congress29.4 %1157
2003Political parties15.7 %1188
2003Judiciary44.2 %1165
2003Police36.9 %1195
2003Armed forces56.4 %1158
2004Congress36.5 %1149
2004Political parties22.8 %1189
2004Judiciary49.7 %1160
2004Police36.3 %1194
2004Armed forces63.0 %1160
2005Congress27.5 %1152
2005Political parties15.8 %1193
2005Judiciary42.7 %1153
2005Police39.0 %1189
2005Armed forces62.6 %1167
2006Congress32.7 %1135
2006Political parties22.7 %1174
2006Judiciary46.1 %1164
2006Police42.5 %1186
2006Armed forces69.0 %1167
2007Congress24.4 %1158
2007Political parties16.1 %1191
2007Judiciary42.0 %1164
2007Police35.8 %1197
2007Armed forces68.5 %1181
2008Congress34.9 %1162
2008Political parties20.7 %1180
2008Judiciary42.8 %1163
2008Police39.6 %1197
2008Armed forces62.6 %1175
2009Congress32.8 %1161
2009Political parties20.8 %1184
2009Judiciary43.0 %1164
2009Police34.4 %1196
2009Armed forces62.8 %1176
2010Congress45.5 %1156
2010Political parties24.7 %1169
2010Judiciary53.0 %1161
2010Police40.7 %1186
2010Armed forces65.9 %1157
2011Congress31.3 %1172
2011Political parties15.9 %1197
2011Judiciary39.8 %1181
2011Police34.8 %1201
2011Armed forces63.6 %1192
2013Congress29.2 %1166
2013Political parties21.6 %1168
2013Judiciary42.9 %1156
2013Police44.2 %1192
2013Armed forces68.9 %1172
2015Congress19.4 %1193
2015Political parties10.4 %1221
2015Judiciary33.2 %1194
2015Police37.1 %1236
2015Armed forces56.7 %1184
2016Congress13.4 %1160
2016Political parties5.3 %1189
2016Judiciary39.9 %1163
2016Police39.5 %1194
2016Armed forces57.4 %1155
2017Congress11.0 %1163
2017Political parties6.7 %1186
2017Judiciary27.8 %1163
2017Police34.5 %1190
2017Armed forces52.4 %1148
2018Congress12.7 %1165
2018Political parties6.3 %1186
2018Judiciary33.7 %1171
2018Police47.6 %1189
2018Armed forces59.6 %1170
2020Congress23.9 %1167
2020Political parties13.1 %1186
2020Judiciary37.0 %1170
2020Police48.2 %1194
2020Armed forces58.0 %1176
2023Congress30.8 %1160
2023Political parties18.6 %1185
2023Judiciary37.5 %1170
2023Police48.2 %1182
2023Armed forces49.7 %1177
2024Congress27.3 %1149
2024Political parties15.9 %1185
2024Judiciary36.6 %1160
2024Police44.7 %1195
2024Armed forces47.8 %1180

The window filters the trajectory and the confidence comparison below. Changes require exact endpoints already present in the research tables.

Compare countries

This comparison always shows confidence levels or their changes, not approval-model coefficients. This comparison includes early partial or incompletely documented frames. It describes covered samples, not an invariant national population.

  • Congress
  • Political parties
  • Judiciary
  • Police
  • Armed forces

Change · Full available history

Compare countries · percentage points
CountryCongressPolitical partiesJudiciaryPoliceArmed forces
Argentina 1995–2024-12.6-9.4-12.3+4.2+16.0
Bolivia 1996–2024-1.9-6.3-12.1-2.3-12.6
Brazil 1995–2024+0.2-1.6-4.2+10.7-13.1
Chile 1995–2024-29.1-21.4-22.3+0.6+5.8
Colombia 1996–2024+1.7-2.6-9.4+20.0+4.5
Costa Rica 1996–2024+3.1-1.8+5.6+24.9—
Dominican Republic 2004–2024-0.6-4.0-12.8+3.3-4.0
Ecuador 1996–2024-9.4-5.9-13.9+4.9-10.7
El Salvador 1996–2024+6.1-6.8+10.8+25.4+26.9
Guatemala 1996–2024-17.0-11.9-16.5-1.5-0.9
Honduras 1996–2024-11.6-12.4-15.1-5.0-9.6
Mexico 1995–2024-8.9-10.0+2.0+0.1+1.5
Nicaragua 1996–2020-13.0-12.4-13.8-14.7-7.1
Panama 1996–2024+4.2-4.8+7.3+17.5—
Paraguay 1995–2024-22.4-9.0-8.0+5.7+13.7
Peru 1995–2024-29.5-12.2-8.3-6.9-15.1
Uruguay 1995–2024+2.5-1.5-4.5+10.8+12.4
Venezuela 1995–2024-2.6+4.4-9.9-2.7-29.4
Change in positive-confidence shares, using exact observed endpoints. Institution-specific denominators. Direction is not a significance test. —: Unavailable. Source: Latinobarómetro / COUNTRY_INSTITUTION_PERIODS.csv.