As public trust is influenced by multiple factors, it is valuable to explore how these factors relate to trust levels simultaneously. This is done through econometric analysis, which examines the links between the outcome of interest – trust in a public institution – and public perceptions of the public governance drivers of trust as well as individual background characteristics. This analysis demonstrates how positive perceptions of specific drivers correlate with the likelihood of placing high or moderately high trust in a public institution, while controlling for perceptions of other drivers and individual background characteristics. Findings from this econometric analysis help identify areas where improvements could significantly boost trust levels.
OECD Survey on Drivers of Trust in Public Institutions in Latin America and the Caribbean 2025 Results
The regression analysis reveals how a positive perception of a public governance driver increases the likelihood of someone having high or moderately high trust in a public institution, while controlling for their assessment of other aspects of public governance and their socio-economic and political background. This approach sheds light on how changing perceptions about specific aspects of public governance (i.e. of a driver) may impact levels of trust. Box A A.1 provides more details on the methodological approach.
However, econometric analysis has limitations and cannot determine whether the drivers cause trust to rise. First, it cannot address reverse causality. For instance, it cannot conclude whether people trust the government more because they believe it uses the best available evidence to make decisions, or whether people believe the government uses the best available evidence to make decisions because they already trust the government. Second, perceptions of different aspects of public governance (i.e. drivers) can be highly correlated. People who believe governments use the best available evidence to make decisions, for instance, also tend to perceive that the government would change course on an unpopular policy (0.56 unweighted correlation). These correlations make it difficult to identify statistically significant effects and may potentially lead to an underestimation of the importance of certain trust “drivers”. Third, factors that are not measured by the OECD Trust Survey can also have an influence on trust, leading to omitted variable bias.
Despite its limitations, econometric analysis remains a valuable tool to identify which public governance aspects are more strongly associated with trust, after accounting for other variables known to affect trust. The results offer governments guidance on which aspects to leverage or improve upon to enhance trust. The choice of the variables included in the analysis is guided by a theoretical framework (Brezzi et al., 2021[1]) that identifies determinants of trust rooted in previous theoretical and empirical research (see Annex 1.A).
The following sections present the results of a logistic regression analysing the likelihood of trust in the national and local governments, the civil service, and legislatures, considering a wide range of explanatory and control variables. The figures show all driver variables with a statistically significant relationship (at the 5% significance level) with trust in the aforementioned institutions. Statistically significant control variables are not displayed, nor are variables with significance level below 5% or with a significance level at 5% and estimated marginal effects under 1.5 percentage points, as their practical policy significance is assessed to be marginal (though these are shown in A.1).
Balancing intergenerational needs and responsiveness to evidence are highly associated with trust in the national government
Two facets of public governance (i.e. drivers) have the highest association with trust in the national government: positive perceptions of government’s capacity to adequately balance the interests of current and future generations and its use of evidence in decision-making. Individuals who are more confident in these two aspects are 7.5 and 7.4 percentage points more likely, respectively, to have high or moderately high trust in the national government.
These two main drivers of trust show different levels of public confidence: 46% of respondents believe their government would fairly balance interests of different generations, while a smaller share (38%) believes that it would use best available evidence in decision-making (Figure A A.1).
Notably, even among those who did not vote for the incumbent government, evidence-based decision making (6.2 percentage points) and balancing the interests of current and future generations (5.6 percentage points) remain the most important drivers of trust in the national government.
Other aspects of public governance also have a significant relationship with trust in the national government. Regarding democratic oversight, confidence in legislatures holding the executive branch of government accountable is associated with a 3.4 percentage point increase in trust. Perceptions of government preparedness for emergencies and regulation of technologies so that the public can use them responsibly are associated with increases of 2.1 and 1.6 percentage points, respectively. Likewise, feeling of having a say in decision-making (1.7 percentage points), government’s responsiveness to inputs in public consultations (2.3 percentage points), and receiving clear explanations about the impact of reforms (1.6 percentage points) are also associated with a moderately increased likelihood of trusting one’s national government.
Just two aspects related to day-to-day interaction between people and the government are correlated with trust in the national government. These are satisfaction with administrative services (2.3 percentage points) and the perception that applications for government benefits would be treated fairly (1.5 percentage points).
Percentage point change in high or moderately high trust in national government in response to a more positive perception of the public governance variables (X-axis) and the unweighted LAC average share of the population with a positive perception of the noted variables (Y-axis)
Note: The figure shows the statistically significant determinants of trust in the national government in a logistic estimation that controls for individual characteristics (age, gender and education level), including whether they voted or would have voted for one of the current parties in power, self-reported levels of interpersonal trust, and country fixed effects. All variables depicted are statistically significant at the p<0.05 level and have an average marginal effect that is equal to or exceeds 1.5 percentage points.
Source: OECD Trust Survey 2023 and OECD-LAC Trust Survey 2025
Satisfaction with administrative services, perceived fairness, and responsiveness are associated with higher trust in the national civil service.
Several aspects of public governance are associated with high or moderately high trust in the national civil service. Satisfaction with administrative services is the variable most strongly associated with this trust. Being more satisfied with administrative services - as 55% of recent users are – increases the likelihood of having high or moderately high trust in the civil service by 4.1 percentage points. Likelihood of trust in the national civil service also increases by 2.9 percentage points when people believe the government bases decisions on the best available evidence, though only 38% of respondents in LAC countries find it likely that the government would do so. Similarly, confidence in governments’ ability to balance intergenerational interests and in legislatures balancing the interests of different groups (which is positively perceived by 34% of respondents) increases the likelihood of high or moderately high trust in the national civil service by 2.9 and 2.7 percentage point, respectively.
Several significant drivers of trust directly linked to people’s day-to-day interactions with public institutions are not particularly well-perceived. These include confidence that public agencies use personal data legitimately (1.9 percentage points), that benefit claims are treated fairly (1.7 percentage points), and that services are improved as a result of complaints (1.8 percentage points).1
Aspects of public governance aspects related to reliability in decision making on complex policy issues, specifically government preparedness for emergencies is associated with a 2.4 percentage point increase in trust likelihood. Integrity, notably the perception that the legislature will hold the executive branch of government accountable, corresponds to a 1.9 percentage points increase in the likelihood of trust in the civil service, respectively. Additionally, belief that the political system is set up in a way that allows people “like you” to have a say in government decision-making is also associated with higher trust in the civil service (2.1 percentage points).2
Percentage point change in high or moderately high trust in the national civil service in response to a more positive perception of the public governance variables (X-axis) and the unweighted LAC average share of the population with a positive perception of the noted variables (Y-axis)
Note: The figure shows the statistically significant determinants of self-reported trust in the national civil service in a logistic estimation that controls for individual characteristics (age, gender and education level), including whether they voted or would have voted for one of the current parties in power, self-reported levels of interpersonal trust, and country fixed effects. All variables depicted are statistically significant at the p<0.05 level and have an average marginal effect that is equal to or exceeds 1.5 percentage points.
Source: OECD Trust Survey 2023 and OECD-LAC Trust Survey 2025.
Trust in legislatures is higher if it is perceived to balance interests of different groups
Latin American legislatures face widespread distrust, with fewer than 30% of citizens reporting moderate to high trust in them. Analysis of OECD Trust Survey results suggest greater confidence in legislatures’ ability to balance the needs of different regions and groups in society shows the strongest correlation with trust in the legislative branch (Figure A A.3.). People who believe legislatures achieve this balance are 7.7 percentage points more likely to trust them, all other factors remaining equal.
The second and third most significant drivers of trust in the legislature are satisfaction with administrative services and feelings of having a say in decision-making, associated with a 3.4 and 2.9 percentage point increase in trust, respectively.
Other factors positively affecting trust in the legislative branch include perceptions of the government fairly balancing interests of different generations (2.6 percentage points), of the legislature effectively overseeing the executive branch (2.5 percentage points), emergency preparedness (2.1 percentage points), responsible use of personal data (1.9 percentage points), and use of quality evidence in decision-making (1.8 percentage points).
Percentage point change in high or moderately high trust in the national legislature in response to a more positive perception of the public governance variables (X-axis) and the unweighted LAC average share of the population with a positive perception of the noted variables (Y-axis)
Note: The figure shows the most robust determinants of self-reported trust in the parliament in a logistic estimation that controls for individual characteristics (age, gender and education level), including whether they voted or would have voted for one of the current parties in power, self-reported levels of interpersonal trust, and country fixed effects. All variables depicted are statistically significant at p<0.05 level and have an average marginal effect that is equal to or exceeds 1.5 percentage points.
Source: OECD Trust Survey 2023 and OECD-LAC Trust Survey 2025.
Providing opportunities to have a say on decisions affecting communities along with satisfaction in administrative services are the most important drivers of trust in the local government.
The two most significant public governance aspects affecting trust in local government relate to perceptions of voice on local issues and to satisfaction with administrative services: people who believe they have opportunities to voice their opinion on decisions affecting their community are 4.7 percentage points more likely to have high or moderately high trust in the local government (Figure A.4).
However, only 38% of Latin Americans believe their local government gives them opportunities to voice their opinions on local matters. Likewise, satisfaction with administrative services is associated with a 4.3 percentage point higher likelihood of high or moderately high trust, while 55% of users reported to be satisfied with these services.
As expected, many aspects related to people’s day-to-day interactions with public institutions also strengthen trust in local government. These include positive perceptions of improving services following complaints (2.4 percentage points), treating claims for benefits fairly (2.3 percentage points), public employees treating everyone fairly (1.9 percentage points), or public agencies use personal data legitimately (1.7 percentage points).3
In addition, aspects related to complex policy making also matter for trust in local government. Preparedness for emergencies is linked to a 3.6 percentage point increase in trust, and legislatures balancing the interests of different groups in society is associated with a 3 percentage point higher likelihood of reporting high to moderately high trust in local government. On both aspects, Latin Americans do not have very positive perceptions (40% and 34%, respectively), highlighting that local governments have a key role to ensure fairness in public policy making across population groups, and in particular, during emergencies.4
Percentage point change in high or moderately high trust in local government in response to a more positive perception of the public governance variables (X-axis) and the unweighted LAC average share of the population with a positive perception of the noted variables (Y-axis)
Note: The figure shows the most robust determinants of self-reported trust in the local government in a logistic estimation that controls for individual characteristics (age, gender and education level), including whether they voted or would have voted for one of the current parties in power, self-reported levels of interpersonal trust, and country fixed effects. All variables depicted are statistically significant at the p<0.05 level and have an average marginal effect that is equal to or exceeds 1.5 percentage points.
Source: OECD Trust Survey 2023 and OECD-LAC Trust Survey 2025.
The econometric results presented below are based on four logistic regression models designed to identify the main drivers of trust in national government, national legislatures, local government, and the civil service across ten Latin American and Caribbean (LAC) countries. Further details on the survey methodology and country coverage can be found in Annex B.
The logistic regression estimates the probability that individual i surveyed in country c expresses trust in a public institution (national government, local government, national legislature, or national civil service). Trust is treated as a binary variable, equal to 1 if the respondent gave a score of 6 or higher on a 0–10 scale (where 0 indicates low trust and 10 high trust), and 0 if the score was between 0 and 4. Neutral responses (score of 5) and “don’t know” answers are excluded from the analysis. The following logit model is applied:
The independent variables used to predict an individual's trust in a public institution are captured by two sets of variables: Driversic and Xic. The Driversic variables refer to public governance aspects considered by the OECD Framework on Drivers of Trust, and reflect respondents’ perceptions of governments’ responsiveness, reliability, openness, integrity, and fairness. The vector Driversic also contains perceptions of government action on intergenerational and global challenges and cultural, economic and political drivers of trust. These variables are expected to be the main determinants of trust across public institutions. In total, 23 Drivers are included, which are listed, together with their estimated coefficients from the logistic model in Table A A.1 and in Annex 1A. To facilitate interpretation and comparability of the regression coefficients, all Driver variables are standardised with a mean 0, and a standard deviation of 1. This standardisation allows the coefficients to be interpreted as the effect of a one standard deviation change in the variable, rather than, for example, a one-unit change. This helps to assess the relative importance of each driver more meaningfully, especially when their distributions differ.
The vector Xic captures demographic, socio-economic and political controls of each respondent that may also impact institutional trust. More specifically, it includes age, age squared, gender, if respondents have gotten a tertiary education or above degree, and level of interpersonal trust (standardised on a scale from 0 to 10) of respondents. They also include dummies indicating if people reported to have financial concerns, or if they are politically aligned with the current government. These controls are included because individuals’ demographic and socio-economic characteristics, as well as their political alignment can influence their perceptions of public institutions (see Chapter 2). In addition, the same set of controls was also used for the regression analysis of the OECD 2024 Trust Survey report (OECD, 2024[2]).
Moreover, all regressions include country fixed effects and survey weights. Missing data are handled through listwise deletion, meaning that any observation with missing values on relevant variables is excluded from the analysis.
In terms of interpretation, all coefficients are shown as marginal effects. For example, the perception that the government makes evidence-based decisions can be interpreted as follows: a one-standard-deviation increase in this perception is associated with a seven-percentage point increase in the likelihood of trusting the national government, holding all other variables constant (0.07 × 100 = 7).
The regression results are largely robust to the choice of model specification. Similar findings are obtained when using an Ordinary Least Squares (OLS) model instead of a logistic regression, and when applying either robust standard errors or clustering standard errors at the country level. Moreover, the magnitude and statistical significance of the coefficients remain broadly consistent when using region (TL1) instead of country fixed effects and by also clustering at this level. In addition, including neutral responses (a score of 5 on the 0-10 scale) among those who indicate low trust (0-4 on the 0-10 scale) or among those who indicate high trust (6-10 on the 0-10 scale) does not substantially change the results – both in terms of magnitude and statistical significance.
The same regression analysis was also replicated solely for the six LAC countries surveyed in 2025 (see appendix B), yielding broadly similar results in terms of the magnitude and significance of the drivers of trust. The 2025 OECD Trust Survey in LAC also included questions on whether respondents had recently experienced adverse events, whether they feared becoming a victim of crime, their confidence in the national government’s ability to cooperate with stakeholders, and their belief that decisions taken by courts are free from political influence. The regression findings indicate that neither recent adverse events nor perceptions of government cooperation with stakeholders have a statistically significant effect on institutional trust. However, not being concerned about becoming victim of a crime is associated with a three-percentage point increase in trust in the national civil service, significant at the 1% level. Similar, believing that courts are free from political influence is linked to a two-percentage point increase in trust in the legislature, also significant at the 1% level
Average marginal effects of selected variable with regard to trust into the national government, national legislature, local government and civil service based on 10 LAC countries
Note: Please note that socio-economic controls refer to the gender, age, age squared, interpersonal trust (standardized on a scale from 0 to 10) of respondents. They also include dummies indicating if respondents have attended tertiary education, have financial concerns and if they support the current government. Significance levels refer to *<0.05, **<0.01 and
***< 0.001. Robust standard errors are applied.
Source: 2023 OECD Trust Survey and 2025 OECD Trust Survey in LAC.
References
[1] Brezzi, M. et al. (2021), “An updated OECD framework on drivers of trust in public institutions to meet current and future challenges”, OECD Working Papers on Public Governance, No. 48, OECD Publishing, Paris, https://doi.org/10.1787/b6c5478c-en.
[2] OECD (2024), OECD Survey on Drivers of Trust in Public Institutions – 2024 Results: Building Trust in a Complex Policy Environment, OECD Publishing, Paris, https://doi.org/10.1787/9a20554b-en.
← 1. There are two further statistically significant public governance perceptions for trust in the national government not mentioned in this enumeration. First, positive perceptions that institutions use data only for legitimate purposes is not listed because its average marginal effect is equal to 1.2 percentage points, below the 1.5 percentage threshold chosen as representing a minimum practical significance. Second, a higher assessed likelihood that people can voice their opinions at the local level is negatively associated with trust in the national government (-2.5 percentage points)
← 2. There is one further statistically significant public governance perception for trust in the national civil service. Confidence in one’s own abilities to participate in politics is associated with a 1.4 percentage points higher likelihood of placing high or moderately high trust in the national civil service, which is below the 1.5 percentage point threshold chosen to assess practical policy relevance in this report.
← 3. The average marginal effect of the perception that civil servants are likely to decline bribes is negatively associated with trust in local government, but the average marginal effect is small (-1.4 percentage points).
← 4. The average marginal effect of the perception that government explains well how people will be affected by a major reform is negative and small (-1.5 percentage points).