2. How trustworthy is your government?

What do people in OECD countries say when asked how much they trust different government institutions? Different institutions and actors elicit different responses. On average across countries, people tend to have high trust in other people. When thinking about government specifically, respondents on average trust their local government more than they trust their national government, and they trust civil servants most of all. Respondents also have fairly high levels of trust in institutions of justice, like the police, courts and the legal system. In contrast, representative legislative institutions, the media and political parties tend to fare the worst – across countries, respondents are most sceptical of these institutions (Figure 2.1).

It is worth noting that awareness of different levels and Ministries in government, as well as their differing responsibilities, can also vary enormously across countries. For this (and other) reasons, Trust Survey questions were adapted to fit local contexts and needs in participating countries, and should be continuously evaluated for cross-national comparability (Box 2.1).

When asked about their degrees of trust in different levels of government, only about four in ten respondents (41.4%) trust their national government, on average across OECD countries, with rates over 50% in Norway,1 Finland, Luxembourg, Ireland and Iceland. 14.8% hold a “neutral” position when evaluating whether they trust their government, and 41.1% tend not to trust their government (Chapter 1).

Local governments tend to inspire more confidence. On average across countries, 46.9% of people say they trust their local government and only 32.4% say they do not trust their local government. Civil servants fare better than the more general local and national governments: half (50.2%) of respondents, on average, say that they trust civil servants in their country. Importantly, fewer than one-third of respondents say that they do not trust civil servants.

However, differences in trust across institutions can also vary widely within countries. For example, 67.6% of respondents in Ireland trust the civil service, while only 50.6% the national government and fewer than half trust the local governments (Figure 2.2). The gap is similar in France.

It should be noted that Japan has high shares of respondents who either feel neutrally about trust in government and civil service or selected “Don’t know,” which is not associated with a number value on the scale. Taken together, a solid majority of respondents in Japan either trust, hold a neutral view, or report they are unsure whether they trust the national government, the local government and civil service. This may suggest an important flexibility in terms of trust in government in Japan and the interpretation of these responses should be explored further (Box 2.1).

The fact that the civil service is viewed as more trustworthy than the more abstract concepts of “national government” and “local government” may be cause for cautious optimism. Civil servants are, in many ways, the human face of government institutions; they work directly and professionally with citizens and users of government services (OECD, 2021[3]). Civil servants are important representatives of government processes and programmes and can be particularly effective and well-perceived when they are autonomous from political influence (Dahlström and Lapuente, 2021[4]). This relatively higher satisfaction with civil servants also aligns with relatively positive perceptions of government reliability (Chapter 4).

Even in countries where trust in the national government was low in cross-national comparison in November 2021, such as Austria – perhaps reflecting the start of the fifth wave of COVID-19 in that country – trust in the civil service remained higher. This suggests some longstanding, structural, underlying confidence in public sector workers.

Public institutions tasked with security and justice also tend to be viewed positively. Over two-thirds (67.1%) of respondents, on average across countries, say that they trust the police. Just over half – 56.9%, on average – trust the courts and legal system (Figure 2.3).

This result roughly aligns with the share of respondents on average who think that courts make decisions free of political influence plus the share who hold a “neutral” view of courts’ independence (Chapter 5). The perceived independence of the courts is positively correlated cross-nationally with public trust in courts and the legal system (Figure 2.4).

It should be noted that the question on “trust in the judiciary and the legal system” may elicit different responses across countries depending on the national organisation of the various functions, and it may be more relevant to further disaggregate these institutions in future iterations of the Trust Survey. The results in Korea, for example, illustrate the possible benefit of better clarifying these institutions: while Korea’s result for trust in the judiciary and the legal system (grouping) is in the lower half of the OECD’s cross-national results, Korea performs well, and above the OECD average, in the more focused question on perceptions of the political independence of the judiciary.

Across countries, one group consistently elicits strong feelings of low trust: political parties. On average only 24.5% of respondents trust political parties, while 55.5% do not trust political parties. Respondents also have relatively weak levels of trust in representative legislative institutions – parliaments and congresses. Only 39.4% of respondents, on average across countries, report trusting their country’s legislative institution. In Denmark, Finland, Ireland, Norway and Luxembourg a small majority do trust their parliament. Indeed, in Norway trust is higher in the parliament than it is in the national government, local government and civil servants.

These results fit into a broader pattern of feelings of disempowerment. Respondents have relatively low levels of confidence in the integrity of elected officials and high shares of people feel their voices are not incorporated in government policy making (Chapter 6). Trust in the national legislature is also strongly influenced by political preferences; while even people who voted for the parties in power do not inherently trust their parliament or congress, people who hold opposing political views exhibit considerably lower levels of trust in their national legislature and in government in general (Chapter 3).

Other institutions do not fare much better in perceptions of trust. Only 38.8% of respondents, on average, say they trust the news media.

These levels of trust in different institutions are driven by governments’ performance in different aspects of governance. The OECD Trust Framework sets out measurable guidelines to estimate where governments are viewed as performing well and where they may be falling short – with direct implications for trust (Chapter 1, Box 1.2).

In nearly every country, respondents are more confident in their government’s reliability than its responsiveness. On average across countries, 47.7% of respondents consider their government reliable and 38.2% say their government is responsive (Figure 2.6). A majority of respondents in half of the surveyed countries (Luxembourg, Denmark, Ireland, Korea, New Zealand, Norway, the Netherlands, Estonia, Iceland, Canada and the United Kingdom) consider their government reliable, as measured by questions on future pandemic preparedness, government use of personal data, and the stability of business conditions. In contrast, in only one country – Korea – do a majority consider their government to be responsive, i.e. responding well to public feedback about policies and services and adopting innovative ideas to improve public services. Estimates of reliability and responsiveness also tend to have a statistically significant relationship with trust in regression analyses, as well (Section 2.4).

An analysis of government values – also defined in the Framework (Chapter 1) – tells a more complicated story. Governments fare best on respondents’ feeling that their own application for a government benefit or service would be treated fairly, one of the dimensions of fairness in the OECD Trust Framework. In general, respondents are sceptical that government “openness” includes real opportunities to engage in the policy-making process – but most feel that they can find information about administrative procedures fairly easily. On average, across countries, 46.2% of respondents consider their government “open”. Perceptions of government integrity are also relatively poor, as evidenced by the average values across questions about petty bribery, revolving doors arrangements for elected and appointed officials, and the political independence of the courts. Only 37.6% of respondents, on average across countries, are confident in the integrity of their government (Figure 2.6).

Interestingly, differences – or the range of results – across countries are relatively low for questions where governments on average scored poorly, such as changing unpopular policies in response to public opinion, using the results of a public consultation, and perception of the likelihood that a high-level political official would refuse a private sector job offer in exchange for a political favour. This means that there is relatively broad agreement, cross-nationally, that governments are not doing well in these areas. In contrast, there is more variation across countries on the questions where governments tended to fare better, on average – on the availability of information on administrative procedures, the legitimate use of personal data, preparedness for a new serious contagious disease, and the fair treatment of applications for public benefits.

Simply put, there is much more agreement among respondents on areas in which governments need to improve, while opinions are more divided on higher-performance areas. This suggests, possibly, a common agenda for OECD countries to address those areas where perception of government performance is widespread low, and benchmark policies and results among countries to continue improve those areas where perceptions are more varied.

Most of the figures in this report present descriptive indicators of public perceptions of different institutions and trust in government. The Trust Survey data are a useful tool for understanding, for example, what share of a national population has confidence in different institutions, services and processes – and for understanding characteristics and perceptions of people who trust (or do not trust) government. This descriptive evidence helps to give a global understanding of the relationship between institutions and trust. 

Understanding the causal relationship between institutions and trust – in other words, how public governance causally affects trust – is a much more complicated task, especially with observational data. Even with the most sophisticated econometrics, the causal relationship between institutions and trust likely moves in two directions. Effective institutions and policies drive trust in government, and trust in government can make institutions and policies more effective. There is also collinearity and interactive effects across different aspects of governance that make it difficult to establish the causal effect of one particular variable. For example, the Trust Survey finds that respondents distrust politicians and are also sceptical of their ability to use their political voice; it is likely that these kinds of variables have an interactive relationship and jointly affect trust.

With these caveats in mind, a simple logit regression analysis of the Trust Survey data presents some evidence of the statistically significant relationship between different institutions and trust in the national government, local government and civil service. Using the pooled cross-national Trust Survey dataset and country fixed effects, we find that different factors are associated with trust in the national government, the civil service or the local government (Box 2.2).

Most of the questions in the Trust Survey can be categorised into the different public governance components of the OECD Trust Framework: reliability, responsiveness, integrity, fairness and openness. Within these, the results on reliability seem to matter most in supporting trust in government.

The use of a regression in the Trust Survey microdata helps us understand the strength and nature (e.g. positive, negative) of the relationship between the dependent variable – trust – and a series of independent variables from the Trust Framework (Box 2.2).

When analysed in a logit regression, all survey questions on reliability have a significant and positive relationship with trust in the national government. For example, holding all other conditions equal, moving from the typical citizen to one with a slightly higher level of confidence in the preparedness to future disease2 is associated with an increase of 6.7 percentage points in the level of trust in the national government. This coefficient, in percentage points, is represented by the blue bar in Figure 2.7 (scale on left y-axis). An increase in people’s confidence on two other “reliability” questions is associated with an increase of around 3 percentage points in trust in the national government (Figure 2.7).

Political drivers, such as the perception of having a say in what the government does, government openness in accounting for views from a public consultation, confidence in the capacity of government to support reforms for the future, and perception of independence of courts, are the other variables with the strongest statistical relationship with trust in the national government.

While these results show how important these factors are vis-à-vis promoting trust, governments face different starting points in how satisfied people are with these different governance factors now. Only 30.2% of respondents, on average cross-nationally, say they feel they have a say in what the government does (right axis in Figure 2.7) – yet this is a fairly important variable related to trust in the national government, as indicated by its relationship with a 5.5 percentage point increase in trust.

Reliability, fairness and responsiveness have the strongest statistically significant relationship with trust in the civil service. Holding all else constant, moving from the typical citizen to one slightly more satisfied with administrative services is associated with an increase of 6 percentage points in the level of trust in the civil service (Figure 2.8, measured by blue bar related to left y-axis). The perception that rich and poor people are treated fairly in applications for public benefits, confidence that the government uses data for legitimate purposes, and confidence in government preparedness for a contagious disease are the other variables most strongly related to trust in the civil service (Figure 2.8).

At the same time, the cross-national average level of satisfaction with the variables shown in yellow vary quite a bit (Figure 2.8). Average values vary from 30.2% of people (cross-nationally) reporting that they can have a say in what the government does to 63% satisfied with administrative services (Figure 2.8, illustrated with the yellow dot related to the right axis). In other words, the starting point in people assessments of government varies across policy dimensions – some policy areas may have a positive and statistically significant relationship with trust, and already benefit from high level of satisfaction (e.g. satisfaction with administrative services). Others are areas that need more improvement.

What influences trust at the local government level? People’s views of government openness and reliability have a statistically significant relationship with trust in the local government. Holding all else constant, moving from the typical citizen to one slightly more confident3 about voicing views on local government decisions or slightly more satisfied with administrative services is associated with an increase of five percentage points in the level of trust in the local government, respectively (Figure 2.9, blue bars associated with the left Y-axis). The other survey questions on reliability (preparedness for future disease, and legitimate use of private data), together with feelings of having a say in what the government does, perceptions that public agencies adopt innovative ideas, and perceptions of equal treatment by public officials, are the other variables with the strongest relationships with trust in local government. At the same time, the starting point in people’s assessment of government varies across policy areas. While a majority of respondents, on average across OECD countries, are satisfied with administrative services (63%) and the use of personal data (51%), only 41% of respondents feel they would be able to voice their views and 30.2% to have a say in what the government does (Figure 2.9 yellow dots, right axis).

These results provide a first exploration of the main factors associated with trust in national government, local government and civil service and show that, on average across countries, these factors vary across institutions. Analysis for specific countries would highlight significant difference within this aggregate picture.


[4] Dahlström, C. and V. Lapuente (2021), “Bureaucracy and Government Quality”, in The Oxford Handbook of the Quality of Government, Oxford University Press, https://doi.org/10.1093/OXFORDHB/9780198858218.013.31.

[1] Moss, F. and B. Vijayendra (2018), When difference doesn’t mean different: Understanding cultural bias in global research studies, Ipsos.

[3] OECD (2021), Public Employment and Management 2021: The Future of the Public Service, OECD Publishing, Paris, https://doi.org/10.1787/938f0d65-en.

[5] Saglie, J. et al. (2021), Lokalvalget 2019: Nye kommuner – nye valg?, Nordic Open Access Scholarly Publishing, https://doi.org/10.23865/NOASP.134.

[2] Yoshino, R. (2015), “Trust of Nations: Looking for More Universal Values for Interpersonal and International Relationships”, Behaviormetrika, Vol. 42/2, pp. 131-166, https://doi.org/10.2333/BHMK.42.131.


← 1. The OECD Trust Survey finds that trust in national government is slightly higher – by about 2 percentage points – than trust in local government in Norway. While it is a very small difference, this stands in contrast to the order of trusted institutions in other countries and in contrast to the results of a Norwegian elections study that measured trust. In this 2019 Norwegian elections study, trust in the municipal council is 5.7 on average – in line with the OECD average result, but higher than trust in the national parliament (5.5) and the national government (5.4) (Saglie et al., 2021[5]). These differences demonstrate that trust levels fluctuate. One potential source of these discrepancies is the timing of the surveys. Trust tends to be higher following elections, which could have influenced the trust averages in the local election study, while the OECD trust survey was fielded during the COVID-19 pandemic.

← 2. In the model this is measured as an increase in one standard deviation.

← 3. In the model this is measured as an increase in one standard deviation.


This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.

Photo credits: Cover picture designed by Together.

Corrigenda to publications may be found on line at: www.oecd.org/about/publishing/corrigenda.htm.

© OECD 2022

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at https://www.oecd.org/termsandconditions.