Building Trust in Government through Citizen Engagement

This paper explores questions related to trust in government and citizen engagement based on a literature review and consultations with experts and World Bank practitioners. The authors unpack the understanding of what trust is by shedding light on different types of trust and their measurements, explore the relationship between trust and citizen engagement, briefly discuss what governments can do to improve their citizens' trust in institutions, and conclude with areas for further exploration. This paper suggests trust in government and citizen engagement form a mutually reinforcing, interdependent dynamic in policy formation that is affected by common attributes and affect development outcomes and effectiveness. It is one in a series of four notes from the Governance Global Practice on citizen engagement in the areas of fragile, conflict, and violence situations, open government, trust, and emerging technology.


Introduction
Citizen 1 engagement -or voice, participation, and accountability -in decision making can help create an inclusive society. An inclusive society must have institutions, structures, and process that empower all groups to participate so they can hold their governments to account, according to the World Bank Group (2014b).
Governments and public institutions play a fundamental role in supporting an inclusive society, yet perception surveys show that citizens' trust in government is diminishing (see figure   1). In 2018, only 43 percent of the global population trusted their governments "to do what is right," according to the Edelman Trust Barometer (Edelman Intelligence 2018). Similarly, the Organisation for Economic Co-operation and Development (OECD) and several international surveys find that trust in government is decreasing (OECD 2017a;Inglehart et al. 2014;Ray 2017). Only 38 percent of people in OECD countries say they trust their government. In 2006, this figure was around 42 percent (OECD 2017b).
Trust in government -or political trust -is a necessary precondition for representative democracy (van der Meer 2017; van der Meer and Zmerli 2017). The erosion of trust in government is thought to indicate "the crisis of democracy" with direct and severe consequences for the quality and ability of representative democracy, its institutions, and its actors (Crozier, Huntington, and Watanuki 1978;van der Meer 2017).
Research continues to claim that erosion of trust in government poses a threat to the quality of representative democracy (Offe 1972;Crozier, Huntington, and Watanuki 1978;Klingemann and Fuchs 1995;Kaase and Newton 1995;Linz and Stepan 1996;Mishler and Rose 1997;Norris 1999Norris , 2011Pharr and Putnam 2000;Dalton 2004;Thomassen 2015). However, little consensus exists in the literature about the definition of trust, the ways to measure it, the cause and implications of the decline of trust in government, and the remedial actions governments can take to restore trust in government and institutions.
Based on a literature review and consultations with experts and World Bank practitioners, this paper explores questions related to trust in government and citizen engagement. It unpacks our understanding of what trust is by shedding light on different types of trust and 1 In the context of the World Bank Group's view of citizen engagement, "the term citizen is not used in a legal sense but is understood in the broad sense of referring to all people in a society or country in an inclusive and nondiscriminatory way" (World Bank Group 2014a). "Citizen engagement" in this paper refers to an inclusive engagement between government as well as public service providers and a wide range of citizens -individual or collective -not bound by a legal sense of citizenship. However, the term citizen may not be understood or applied in a similar manner when it comes to measuring trust in governments because of differences in methodology. Given the measurement of political trust heavily relies on surveys, the representative samples of a country or area may be based on citizenship, which the authors are unable to determine from publicly available information. their measurements. Then, it explores the relationship between trust and citizen engagement, which the authors argue are part of a mutually reinforcing dynamic. The last section briefly discusses what governments can do to improve their citizens' trust in institutions and concludes with areas for further exploration. Note: The graph compares changes in the country scores on the Edelman's Trust Barometer from 2011 (x-axis) to 2017 (y-axis). Any country with an increase in trust in institutions will be above the 45-degree line (and shaded green), while those with a decrease in trust will be below the 45-degree line (and

Unpacking Trust
To understand the claim of declining trust in government, it is important to first understand the concept of trust and how trust is being measured. This section focuses on two questions: What is trust, and how is it measured?

Trust -An Ambiguous Concept
Scholars define trust in various ways, using relational and situational elements and a combination of them. Trust is referred to as the underpinning of all human contacts and institutional interactions and can be defined as the willingness of one party to rely on the other party to keep its commitments (Blind 2007;Tonkiss et al. 2000;Misztal 1996). Uslaner (2002) argues that trust should be explained by relationship, as simple as "A trusts," to transcend the interpersonal trust beyond one's own experiences, grouping, or associations including friendship. Other scholars argue that the broad notion of interpersonal trust can be explained with two distinctive kinds -"limited" trust, which represents trust between two people who know each other well (e.g., family, friends, and neighbors), and "generalized" trust, which refers to trust between casual acquaintances and complete strangers (Dehley, Newton, and Welzel 2011;Murtin et al. 2018;Putnam 2000). Van der Meer and Zmerli (2017) argue that trust -particularly political trust -is fundamentally relational and situational. Hardin (2002) explains trust as "A trusts B to do X." Bauer adds the concept of time to Hardin's definition to indicate that trust may change over time -"A trusts B to do X at T" (Bauer and Freitag 2017).
For the purpose of this paper, the distinction between social and political trust is important.
Social and political trust are conceptually distinct (see table 1) though they reinforce to each other. Uslaner's definition of generalized trust "A trusts" may conceptually better explain social trust, which is "trust in people" in general for a nonspecific purpose or situation, while Hardin's "A trusts B to do X" better explains political trust as one's confidence in institutions or actors for a specific purpose. Political trust can change over time, reflecting the shortterm outcomes, evaluations of particular leaders and institutions, and the public's expectations toward their governments and institutions. The difference between the two essentially is regarding the specification of both trustee and expected behavior (Bauer and Freitag 2017).
As Norris (2011) argues, it is natural to think that the expectation toward fellow citizens, families, friends, or neighbors (i.e., social trust) differs from that toward governments and institutions (i.e., political trust). This paper focuses on political trust. Stems from socialization through one's parents.
Social trust is generalized trust, which is not directed at specific people for a specific purpose. It is trust, confidence, or faith in strangers. Therefore, the concept of social trust is beyond the interpersonal concept of "us versus them" that is based on experiences, grouping (e.g., education, ethnicity, culture, gender, geographical locations, and income level), and associations (e.g., friendship, kinship, and neighbors).
Social trust reflects long-term optimism.

Political
Political trust is confidence in institutions and actors (e.g., the executive, legislative, judiciary, the bureaucracy, police, the media, private sector or business, nongovernmental organizations, and regional or international organizations).
Stems from group membership, government policies, or the general labels of political support and satisfaction.
Political trust is particularized trust, which is based on ties to one's own in-group and political trust toward specific institutions and actors often for a specific purpose and situation.
Related concepts to political trust include political efficacy, which is defined by office-holders' responsiveness to their constituency and citizens (i.e., accountability), and political cynicism, which is the negative evaluation of the inherent nature of politics.
Trust in government reflects short-term outcomes, the evaluations of particular leaders and institutions (e.g., ability to deliver quality public services, respond to citizens' needs and demands, and to effectively manage social, economic, and political uncertainties), and expectations.
Political trust is government performance minus expectations.
The relationship between social and political trust remains a topic to be examined further.
While scholars such as Uslaner (2017b) argue there is little relationship between trust in people and confidence in institutions or actors, some surveys find only a weak correlation between the two (Newton 2001). Yet it is intuitive to assume these are interrelated for various reasons (van der Meer 2017; Zmerli and Newton 2017). Zmerli and Newton (2017) for example argue that social trust is a prerequisite for political trust, placing social and political trust in hierarchical relationship. Braithwaite (1998) sees they have a mutually constituting relationship. People may trust an institution because they trust its agents, and they might trust the institution's agents because they trust the institution. Furthermore, several scholars argue that positive personal experiences among individuals build interpersonal trust and will have a "spill-over effect" onto governmental trust (van Ingen and Bekkers 2015). While some echo this view, many scholars disagree with lumping together social and political trust.

Taking a Closer Look at Political Trust
This paper does not distinguish trust in competence (e.g., whether institutions' actions are matching people's expectations of their competency) and trust in intentions (e.g., whether institutions are perceived as ethical and fair) (Nooteboom 2007;OECD 2017b). Trust in government and confidence in government, for example, might mean different things to different people. "Confidence" connotates competence, while "trust" resonates to integrity, and it is a value. Although important, these distinctions are beyond the scope of this paper, and its authors will often use "confidence" and "trust" interchangeably to unpack political trust.
Decline of trust in government is explained by multiple factors. Researchers find strong correlation between trust erosion and corruption (Braithwaite 1998;van der Meer 2010;van der Meer and Zmerli 2017;Newton, Stolle, and Zmerli 2017;Uslaner 2017a,b). Some argue performance and quality of government -including a failure to reduce economic inequal-  Dalton (2004) argues the decrease in trust in government is to be blamed on changing citizen values and expectations. Both political and personal expectations (e.g., past experiences, perceived levels of risks such as unemployment, and financial and economic volatility) shape people's trust in governments and institutions (Dalton 2004;Murtin et al. 2018). The main correlation studied in the literature is between trust in government and citizens' willingness to comply with laws, regulations, and tax demands. There is evidence that low citizen trust in government can weaken the social contract and lead to the disengagement of citizens and firms from the state in several key dimensions (Arizti et al. 2010).
Not only government performance and expectations, but also people's experience in the process of government decision making influence the level of trust in government. Political trust is often understood as people's trust in their governments based on performance and expectations. The simple formula of this concept is "government performance -expectations = political trust." This type of political trust is known as "outcome-based" trust. It is formed regardless of citizen engagement or interactions with government in the decision-making process. At the same time, people's engagement experience with government can also influence the level of political trust. This type is known as "process-based" trust and is determined by the citizens' satisfaction on the level, depth, and quality of engagement (e.g., engagement type, frequency, and responsiveness) in the government's decision-making process. People may or may not be satisfied with an engagement's outcome, but the engagement in the process may allow government to earn legitimacy and the confidence of citizens.
For example, a result of consultations may not be in people's favor, but people consulted may be satisfied with the level, depth, or quality of consultations. This type of trust is described as "process-based" trust in government.
Some authors also argue that while some trust in government is good, a lot is not always better. There is some evidence that there can be "too much" trust in government. Presumably many of the highest-trust respondents in survey questions are excessively deferential to authority. If government legitimacy rests on its programs and policies and not exclusively on political and cultural associations, skepticism about government is healthy and natural (Cook and Gronke 2005;Aritzi et al. 2010). Given the existence of political opportunism, some level of distrust in the government is fundamental (North and Weingast 1989).
While an optimum level of trust is important for governments to function well, some scholars have found that trust in government in closed polities is antithetical to representative democracy. A recent body of literature highlighted the link between political dissatisfaction and growing democratic awareness. The "critical citizens," as Norris (1999) defines them, are individuals that support democracy and are skeptical of political leaders. As people converge toward more democratic awareness, they demand that governments be held accountable, and skepticism rises (Jamal 2007). Jamal (2007) looks at the levels of trust in the Arab world. The author argues that contrary to what happens in the Western world, individuals who support democracy will hold lower levels of political trust in the closed polities of the Arab world. On the other hand, citizens happy with the status quo and the levels of human rights have lower support for democracy and higher levels of political confidence in their leaders.
Several political philosophers such as Locke, Hume, Montesquieu, and Madison argue governments should be distrusted by their citizens, and institutions should be built in a way to weather citizen distrust of the government (Hardin 2002). In the view of these liberal theorists, the only intelligent stance citizens should take toward their governments is distrust. Cook, Hardin, and Levi (2005) discuss the importance of distrust and of strong institutions that can operate successfully "despite pervasive and active distrust." The authors argue that contrary to what others believe, the fall in trust in some democracies is not harmful or detrimental to democracy.
Some scholars argue it is important to live in a culture that values and nurtures interpersonal or social trust. Social trust is empirically associated with measures of good government, indicating that trust between citizens improves social cohesion and reduces social conflicts (Newton, Stolle, and Zmerli 2017), while such relationship has not been established empirically for political trust. Other academic research shows that confidence in institutions increases efficiency and reduces costs (Fukuyama 1995;Putnam 1993;Zak and Knack 2001).
Social trust builds enough confidence in institutions to enable those same institutions to create social wealth and order. When interpersonal trust is scarce or limited to an ethnic group or circle, it prevents the effective and efficient rule of law. In other words, there needs to be enough trust in institutions to assure these benefits, and maximizing interpersonal trust might be the way of achieving it (Braithwaite 1998).
The political trust research essentially seems to suggest the following: (i) complex contrib- understanding of what trust is, it is therefore important to look at how trust is measured as many of these claims are based on the analysis and interpretation of trust data.

Measurements of Trust
The measurement of trust is a contested field.
Though each trust measurement is disputed, trust has been measured in three ways: (i) subjective and direct through perception surveys; (ii) objective and indirect through objective data used as proxies (Kucher and Götte 1998); and (iii) through experimental measures of trust using behavioral games (Glaeser et al. 2000;Habyarimana et al. 2009). Scholars can use perception surveys to ask direct questions about trust in government. They can also use experimental games and proxies to observe participants' behavior, which are indirect yet objective. Though these methods are objective, usage of attitudinal proxies as measurements of trust in government -such as the voluntary act of casting votes in elections, participating in political activities, paying tax, and obeying the law -is perceived as problematic by some scholars given these proxies are behavioral outcomes of, arguably, trust in government. As Norris (2017) indicates, drawing a direct causal relationship between trust in government and an individual's behavior, say toward tax compliance, would be difficult as there may be various reasons for an individual to demonstrate a particular attitude. Experimental games often are used to measure general or social trust.
The following discussion of two methods -perception surveys and proxies -intends to deepen an understanding of how political trust is measured.

PERCEPTION SURVEYS
Perception surveys are the most popular method to measure political trust to date. Data analysis of such perception surveys -international, regional, or national -have been used by researchers to make various claims, including trust erosion and its potential consequences.
Yet, the various measures of trust in government that result from surveys are at the center of controversy in the academic world. Trust-related terms such as "confidence," "performance," and "legitimacy" are often used interchangeably as "trust" though they have different connotations. Table A.1 in appendix A gives examples of international, regional, and national surveys that collect data on citizens' political trust.
Despite variation in framing, the essence of political trust survey questions seems to stem from the relational concept of political trust -"Do you (A) trust B to do X?" -and includes a broader, more generalized measurement of social trust, such as "Do you (A) trust B?" As discussed previously, conceptually, scholars make distinctions between social and political trust and argue their correlations. When it comes to the practice of measuring political trust, these conceptual distinctions are not necessarily reflected, leaving room for respondents' personal interpretations of the questions.
The key concepts measured by perception surveys vary (Arizti et al. 2010). Government performance is usually an empirically observed pattern of past government behavior. Trustworthiness of government is usually a predicted estimation of how government and its institutions are likely to act in the future. Trust in government generally is an assessment of the government's ability to deliver on its promises. Perception surveys are considered good ways to measure trust as long as their limitations are recognized. As with all intan-gible concepts, measuring trust raises several questions about interpretation and subjective judgment. For instance, some authors argue low levels of trust represent a profound alienation (Abramson 1983) while others believe that decreased levels of trust represent a superficial disapproval with a present administration (Cook and Gronke 2005). Perception surveys measuring institutional trust are believed to perform worse than those that look at interpersonal trust (OECD 2017b).
The survey method of measuring trust has two major drawbacks. First, when analyzing survey data, it is difficult to determine whether the coefficients measure what they are supposed to measure and whether trust is correlated with other -possibly omitted -determinants of legitimacy, compliance, and so on. Second, surveys do not measure actual behavior, but mere intentions (Kucher and Götte 1998). Furthermore, Eichenberger and Oberholzer-Gee (1998) offer evidence through behavioral experiments that moral behavior is more frequent when the cost of moral attitude and actions is low. Therefore, experimental subjects showed the highest developed morals in survey situations because the cost of answering morally was zero. For these reasons, the authors of this paper are cautious about overreliance on survey data on trust and their interpretation.
International and regional measures used to assess trust have significant shortcomings linked to their methodology and analysis. Survey coverage is often uneven both across countries and over time. Most survey data come from a small sample (i.e., 1,000 to 1,500 respondents per country), precluding intracountry analysis and failing to be representative of the whole population. It is for these reasons the OECD (2013a) recommends that currently available surveys not be used to support policy analysis or lead to policy recommendations.
Different surveys have specific strengths to answer different research questions. For example, the Gallup World Poll has the largest coverage since 2009, and it might be the best dataset for comparisons across countries and over short periods of time or for studying relationships between trust and other country-level variables from other data sources. On the other hand, for studying relationships between trust and other respondent characteristics, either the Afrobarometer or the World Values Survey (WVS) is more appropriate because each is much richer and more detailed.
Differences in wording of trust questions need to be interpreted carefully, as it might in- fluence how respondents answer. The Afrobarometer, for example, asks about how much "trust" respondents have in government. The WVS asks about "confidence" in a list of "organizations" including government. In this case, "confidence" may connote competence dimensions more, while "trust" may instead relate more to integrity and value dimensions.
Cultural aspects, such as language, of the survey responders and methods need to be taken into consideration. The Afrobarometer measure of trust, as far as the formation of questions go, is one of the more accurate among international surveys. It asks more specific questions about government performance in various sectors or policy areas. However, the standard Afrobarometer survey instrument collects information only in the interviewer's native language, which may not be the survey respondents' native language. In fact, only 36 percent of survey respondents spoke the native language of the surveyor (Adida et al. 2016). Asunka (2015) argues survey responders are also sensitive to who they believe sent the interviewer to conduct the survey, whether other people are present during the interview, and the interviewer's gender.
In addition to the perception survey methods, the interpretation and analysis of such perception data influence the outcomes of the trust measurements. Cross-national comparison of trust data itself may not tell the whole story of trust surveys.
National averages of trust in government rank between 80 percent in Switzerland to 12 percent in Greece (OECD 2013a), and they do not systematically reflect standard of living, gross domestic product, or growth levels. For example, according to the Gallup World Poll, Turkey, a middle-income country, has much higher trust levels than Japan or the United States.
Different levels of confidence in the government can be explained not only by cultural and context-specific factors, but also by the different expectations citizens have of government services and performance based on stages of socioeconomic development (OECD 2013b). For these reasons, trust in government might not respond to long-term economic development or standard of living, but more to cultural factors, expectations, and political environment. In Middle Eastern countries, for example, what matters most is not government performance, but personal relationships and social ties (Brixi et al. 2015).
National surveys are usually not comparable across countries and are often collected on an adhoc basis only, but they can better inform policy decisions than international ones. National surveys provide greater insights into the drivers of trust and can account for election cycles. They tend to cover trust questions more in detail and can offer country-specific insights and measures of trust into existing country policies and policy implementation (OECD 2013b). For example, the American National Election Studies (ANES) asks more specifically: "How much of the time do you think you can trust the government to do what is right?" This wording may connote integrity and fairness dimensions relative to competency dimensions. The ANES also asks more specific trust-related questions such as: "Would you say the government is pretty much run by a few big interests looking out for themselves or that it is run for the benefit of all the people?" "Do you think that people in the government waste a lot of money we pay in taxes, waste some of it, or don't waste very much of it?" "Do you think that quite a few of the people running the govern-

ATTITUDINAL PROXIES AND OBJECTIVE DATA
Trust is also hard to measure with objective and indirect data because a good proxy for trust must reflect whether individuals believe that the government's behavior is in their best interest or not. The Kucher and Götte (1998) study looked at whether trust in institutions is a determinant of tax compliance. It may be the only work that uses objective measures as proxies for trust. Instead of using perception surveys to measure trust and confidence in government, the authors use a ratio of government suggestions and actual results in public votes. The authors argue the higher the concurrence of suggestions and results, the less opportunistic is a government and, therefore, the higher is people's trust in its authority.
Brinkerhoff, Weterberg, and Dunn (2012) use willingness to pay for improvements in the water service as a proxy for users' trust in government. This approach is an extension of previous work that relied on tax compliance and user fees as indicators of citizens' trust in government (Glaser and Hildreth 1999;Fjeldstad 2004).
Several scholars have looked at the correlation between trust measures and tax compliance apart from the examples cited above (Feld and Frey 2002;Aritzi et. al. 2010). Several surveys include tax compliance or evasion questions to calculate trust indicators. Using survey results from the WVS and the European Value Study, Chan, Supriyadi, and Torgler (2018) find statistically significant positive correlation between trust and tax morale. The authors distinguish between horizontal (generalized) and vertical (institutional) trust. They find no correlation between the first and tax compliance but find a strong correlation between tax morale and institutional trust -or trust in government. Thus, according to them, vertical trust is a key factor in understanding tax morale across different cultures and institutions, and it matters more than horizontal trust.
On the other hand, Norris (2017) (Braithwaite 1998). For this reason, the empirical evidence of the relationship between citizen engagement and both types of trust is explored.
Few studies look at the impact of trust in general -and more specifically, the consequences of political trust -and they tend to have a narrow scope. For example, some scholars have studied the relationship of trust and civic engagement and participation. In these studies, participation refers to elections, which is an important form of citizen participation in decision making yet limited in scope to measure "participation" in an inclusive society. Citizen engagement in decision making referred in this paper can take place through elections, political and social organizations, and other forms of direct participation and deliberation (World Bank 2017).
Citizen engagement offers a way to (re)build and enhance trust, according to several scholars of social capital theory (Brehm and Rahn 1997;Claibourn and Martin 2000;Jennings and Stoker 2004;Paxton 2007). This notion was also reflected in an OECD civil society organi- The literature agrees it is difficult to establish a causal relationship between citizen engagement and trust in government, but most authors follow Putnam who states, "civic engagement and trust are mutually reinforcing" (2000,137). In other words, without citizens' trust in government, formal citizen engagement is unlikely. Without citizens' participation, government's performance will be poor, and trust in government will fall (Brixi et al. 2015). Low trust in public institutions is part of the reason why citizens do not engage, and the lack of citizen participation in government decision making negatively affects performance and accountability, which leads to a decrease in trust.
Based on our knowledge in trust research to date, this paper suggests trust in government and citizen engagement form a mutually reinforcing, interdependent dynamic in the policy arena 2 that is affected by common attributes and affects development outcomes and effectiveness. Inclusive citizen engagement is an approach for state-citizen interactions in a policy arena to provide citizens a stake in decision making. When citizen engagement is designed and implemented well, it provides government an opportunity to foster "process-based" trust in public deliberation and service delivery. While trust may be one determinant for citizens to participate in this process, citizens' experiences and satisfaction in the process could also shape trust in government. Country contexts, roughly represented by terms and notions -such as the norms and culture, the rules, and the availability of an enabling environment through, for example, access to information laws, freedom of expression and association, and anti-corruption laws -seem to influence both trust in government and citizen engagement. Demographics would equally affect citizen engagement, which calls for inclusive and equitable outreach efforts for participation.

Earning Trust
Trust research on how to earn or (re)build trust is limited but offers some pointers for action.
The earlier sections of this paper summarize the factors that correlate strongly with erosion of political trust, including corruption, performance, and quality of government; exposure to violence; increases in citizen expectations; and the state of the economy at the time of the survey. Although it is not known whether these factors also contribute to earning or (re) building political trust, it may be assumed that improvements in these areas could contribute to avoiding further erosion of political trust. In the realm of governance, defined as "the process through which state and nonstate actors interact to design and implement policies within a given set of formal and informal rules that shape and are shaped by power" (World Bank 2017), corruption as well as performance and quality of government may be two areas governments could address to boost levels of both outcome and process-based trust. They are tightly interlinked areas. One could argue the corruption level is part of government performance and quality.
Another way to discuss what governments can do to earn trust is formed around the level of government functions and engagement. For example, Kettl (2017) introduced the wholesale and retail dimensions for governments to earn citizens' trust. He argued the wholesale level is for governments to earn trust through policy formulation -or in other words, at

Photo by Mohammad Al-Arief/The World Bank
the macro level of government and citizen interactions -while the retail level relates to the governments' ability to implement its policies at the micro-level, say through day-today interactions with citizens to deliver public services where governments could improve citizens' experiences with them. Kettl (2017) also pointed out that transparency is a critical foundation for trust in government. However, a complex concept of transparency itself can be as difficult an issue because it could potentially undermine trust by overflooding citizens with data and information.
Empirical data on how a participatory approach to government decision making (i.e., citizen engagement) affects the level of trust in government are limited. The literature seems to indicate that citizen engagement, which enhances transparency and accountability, helps to build legitimacy and trust in government. Open, equitable, and inclusive policy making is most often promoted as a means of improving democratic performance and efficient and effective administration (Shah 2007).
The relevance of citizen engagement in decision making throughout the public policy process has been extensively studied in recent years. Scholars argue public participation has a positive effect on the participants' satisfaction with political outcomes and the legitimacy of procedure (Frey and Stutzer 2005). The literature also supports the idea that citizen engagement is fundamental for citizens' satisfaction in situations where a certain amount of conflict is involved (Traber 2013). Citizen participation throughout the public policy process is also key to achieve the internationally agreed-on development goals, including the Sustainable Development Goals (Bertucci 2007).
Citizen engagement enables citizens to hold government and service providers to account by giving them platforms for voice and agency. Its potential outcome areas are wide, including benefits in the state (e.g., better governance), in state and society relationships (e.g., improved legitimacy), and in society (e.g., improved allocation of resources and provision of public goods) as well as instrumental benefits (e.g., improved allocation of resources and provision of public goods) and institutional benefits (e.g., inclusive state building).
Citizen engagement, however, needs to be carefully designed and implemented, as widely acknowledged in social accountability research (Grandvoinnet, Aslam, and Raha 2015). Critics of citizen engagement point out that the process could potentially delay decisions, increase conflict, and disappoint participants when expectations are not met, and lead to a downward spiral of reduced accountability and more distrust. In today's post-Arab Spring world, the value of citizen engagement perhaps outweighs the risk and cost of not engaging citizens in deliberative process. Understanding the context for citizen engagement -the state and relationship of state and citizen actions; state and relationship of citizen engagement levers such as informa-tion, citizen-state interface, and civic mobilization; and contextual factors that can influence these elements in the deliberative policy arena (see figure 2) -becomes critical.

State Action
• Awareness of the issue Various participatory approaches are sought to improve transparency and accountability at the micro-level. For example: Dialogue between governments and citizens is a useful mechanism for increasing ac- Social media platforms in the past decade, for example, quickly became tools for instantaneous, direct, one-on-one interaction interfaces for governments and citizens. They also became platforms to generate news and information, which may or may not be accurate. This paper does not discuss the roles social media plays in advancing or shifting an understanding of transparency or how it affects the level or status of trust. These areas may be ripe for further study. Additionally, whether citizens will trust social media and other information from the internet may be another emerging question to be explored.
Citizen engagement in service delivery helps governments achieve better development outcomes by improving programming effectiveness, detecting and reducing corruption, increasing awareness about services, and reducing costs. A report by the OECD (2011) states that citizen engagement in service delivery can foster a more efficient and effective use of resources by reducing costs to the government while increasing user satisfaction (Entwistle and Martin 2005). Decentralization, which brings governments closer to citizens, and citizen engagement in decentralized service delivery offer opportunities for transparent, citizen-centric, and accountable governance, possibly leading to improved trust in institutions among citizens. Citizen participation in service delivery can also help tackle service failures, identify solutions to complex problems, and contribute to enhancing societal well-being. Furthermore, public inclusion can improve democratic governance and build public trust (OECD 2011).
Improving elements related to trust in tax systems can potentially improve trust in gov-ernment by improving customer experience. The World Bank's research on innovations in tax compliance identifies four key drivers of trust in tax systems to shape tax reforms as an enabler of trust in government. These drivers include fairness, equity, reciprocity, and accountability. This research captures the extent to which (i) tax systems are fairly and competently designed and administered ("fairness"); (ii) burdens are equitably distributed and everyone pays their share ("equity"); (iii) tax revenues are translated into reciprocal publicly provided goods and services ("reciprocity"); and (iv) governments administering those tax systems are accountable to taxpayers ("accountability"). Research also defines fairness and equity as tax system outcomes and reciprocity and accountability as tax governance outcomes. Together, these two categories of outcomes, which constitute trust in tax systems, affect citizens' tax compliance, which may also be influenced by norms, values, and ethics.

Conclusion
This paper provides a review of what trust is, how it is measured, the relationship between trust and citizen engagement, and what governments can do to improve their citizens' trust.
The trust research universe is diverse and complex. Research is limited in amount and scope to exploring the relationship between trust and citizen engagement and what governments can do to improve trust. The availability of data to verify the claim of trust erosion and governance is limited due to the challenges of trust measurements and their interpretation.
This paper recognizes the potential room for discovering ways to generate more specific political trust data by refining survey questions. They could include those that would explore linking government quality, performance, expectations, and corruption with the level of trust in institutions, government, and other state actors. Further analysis, data collection, and impact evaluation -and a framework or methodology to design impact evaluations in the trust area -can be useful to fill the knowledge gaps that exist on government performance in public deliberation, service delivery, transparency and accountability, and trust in institutions.

Photo by Sarah Farhat / World Bank
Understanding trust in different contexts, for example, in fragile and conflict-affected environments or closed polities, is another area for more exploration. As government systems continue to be digitized and digital connectivity advances, the internet of things and other emerging technologies (e.g., artificial intelligence) could potentially support citizen and government interactions in public deliberation and service delivery. In today's environment, trust research should also consider studying how GovTech and CivicTech influence the level of trust in government, how transparency enabled by technology influences trust, and whether citizens will trust emerging technologies, such as artificial intelligence, in public participation arenas.   Institutions: government, business, media, and nongovernmental organizations (NGOs).

Examples of Political Trust Surveys
Answer Scale 9-point scale: 1 means "do not trust at all" and 9 means "trust them a great deal." Answer Scale 2 answer options: approve/ disapprove: yes or no.

Information
At least 1,000 individuals. In some large countries, such as China and the Russian Federation, sample sizes of at least 2,000 are collected. Although rare, in some instances, the sample size is between 500 and 1,000. Surveys are based on telephones in countries where coverage represents at least 80 percent of the population or is the customary survey methodology.
1,150 respondents per country for general online population.

Level Global
Survey Institutional Profiles Database (IPD)

# of Countries 144 countries in 2016 survey (remains in progress)
Years 2001, 2006, 2009, 2012, and 2016 (in progress) Frequency Irregular waves Measurement Trust and Legitimacy: Countries' institutional characteristics of nine functions and four sectors.
Functions related to political trust include political institutions; security, law and order, control of violence, functioning of public administrations, and openness.
Sectors include public institutions, civil society; markets for goods and services; the capital market; and the labor market and social relations.
Does the legitimacy of the political authorities stem from their ability to ensure for large sections of the population: (i) an economic benefit (jobs, income); (ii) a social benefit (health, culture); or (iii) a sense of national pride?

Methodology Information
The 2012   Answer Scale 4-point scale: 1 means "a great deal," 2 means "quite a lot," 3 means "not very much" and 4 "none at all."

Methodology Information
The number of completed interviews included in the national dataset in most countries is 1,200. Survey data collection is based on face-to-face interviews at respondent's home or place of preference. Answer Scale 4-point scale: 0 means "not at all," 1 means "just a little," 2 means "somewhat," and 3 means "a lot" with an option of answering "don't know or haven't heard."

Methodology Information
Face-to-face interviews with a randomly selected sample of 1,200 or 2,400 people in each country. Interviews usually take 1 hour to cover 100 questions. Each survey round usually takes about 12 months to survey all countries. A master questionnaire is provided in English, French, or Portuguese. In the Round 7 questionnaire, five spaces are included for country-specific questions.

Level Regional
Survey Americas-Barometer

# of Countries 34 countries
Years 7  Answer Scale 7-point scale: 1 means "not at all" and 7 means "a lot" with options of "don't know" and "no answer." 4-point scale: 1 means "very trustworthy," 2 means "somewhat trustworthy," 3 means "not very trustworthy," and 4 means "not at all trustworthy" with options of "don't know or no opinion," "no answer," and "inapplicable."

Methodology Information
Face-to-face interviews with samples of 1,500 respondents per country.

AmericasBarometer includes interviews in indigenous languages (15 total).
AmericasBarometer is housed under the Vanderbilt University's Latin American Public Opinion Project.

Level Regional
Survey Arab Barometer

# of Countries 14 countries
Years 4  "How much trust do you have in the following?" Arab nationalist parties, socialist or leftist parties, Islamist parties, and nationalist parties.
Answer Scale 4-point scale: 1 means "a great deal of trust," 2 means "quite a lot of trust," 3 means "not very much trust," and 4 means "no trust at all" with options of "I don't know" and "declined to answer."

Methodology Information
Face-to-face interviews in the respondent's residence. All country surveys are based on probability samples representative of citizen aged 18 or above, which translates into approximately 1,200 respondents on average.

Level Regional
Survey Asian Barometer The president or prime minister, courts, national government, political parties, parliament, civil service, military or armed forces, police, local government, newspapers, television, the election commission, and NGOs.
Answer Scale 4-point scale: 1 means "a great deal of trust," 2 means "quite a lot of trust," 3 means "not very much trust," and 4 means "none at all" with options of "do not understand the question," "can't choose," and "decline to answer."

Methodology Information
Face-to-face interviews in respondents' homes or workplace in the language of the respondent's choice. A model Asian Barometer Survey has a sample size of 1,200 respondents per country. "I would like to ask you a question about how much trust you have in certain media and institutions. For each of the following media and institutions, please tell me if you tend to trust it or tend not to trust it." The media, political parties, justice and the (NATIONALITY) legal system, police, army, public administration in (YOUR COUNTRY), regional or local public authorities, the (NATIONALITY) government, (NATIONALITY) parliament, European Union, United Nations.
"And please tell me if you tend to trust or tend not to trust these European institutions." Answer Scale 4 answer options: "great deal," "fair amount," "not very much," and "none at all" with an option of answering "no opinion."

Methodology Information
A random sampling of about 1,000 adults aged 18 and older living in all 50 U.S. states and the District of Columbia. The polls are based on telephone interviews.

Level National
Survey Organisation for Economic Co-operation and Development (OECD) Trustlab Self-reported trust in government: "When answering the following questions, please think about [enter country here] institutions. How much trust do you have in the government?
Self-reported trust in judicial systems: "How much trust do you have in the judicial system?" Answer Scale 11-point scale: 11 discrete options ranging from 0 (I don't trust them at all) to 10 (I completely trust them).

Information
There are three modules in the platform. Module 1 contains three behavioral games to capture measures of social norms. Module 2 is an Implicit Association Test a to capture implicit levels of trust in government and in the judicial system. Module 3 is a traditional survey module with an extensive set of questions on interpersonal and institutional trust.
Country-specific modules can be included.
A minimum sample of 1,000 respondents in each country. The sample is provided by a private sector polling company and is nationally representative by age, gender, and income. Participants complete the platform online using a link provided by the polling company. Upon completion of the entire platform, participants are rewarded through a lump sum of their time and with an additional payoff that they can earn in the behavioral game. Certain exclusion criteria are applied to filter out poor quality responses according to quality assurance measures.
a. The Implicit Association Test (IAT) is an experimental measure of trust in institutions that asks respondents to rapidly sort relevant words to the left-and right-hand sides of the computer screen. The IAT relies on the idea that a person will react more quickly when the concept and the evaluation that one makes of this concept