The Relationship between Ombudsman, Government, and Citizens: A Survey Analysis

The first modern ombudsman's office was established in Sweden in 1809. Its task was to protect the rights of citizens against the executive branch. Since the 1960s, the ombudsman concept seems to have traveled around the world. Ombudsman institutions currently exist on at least one government level in most countries in the world. For citizens, ombudsmen are an alternative to expensive and complicated judicial procedures. Individual problems are often solved in a quick and flexible way. This is the individual role of ombudsmen.

Based on their experience with citizens' complaints, ombudsmen give recommendations that seek to alter laws, regulations, and/or organizational structures. This is the collective dimension of the ombudsman function. The ombudsman does not have the power to make binding decisions but does have the right to reveal problems within organizations and persuade those organizations to follow his or her recommendations (Reif 1999).

The role of the public sector ombudsman has evolved in different ways in different countries. Roy Gregory (2002) distinguishes between two main types of public sector ombudsmen, depending on the country they function in. The pioneers of ombudsmanship (Sweden, Finland, Norway, Denmark, and New Zealand) are all high‐income, stable democracies with rule of law. The objective of these ombudsman services has been to provide a fast, effective, and user‐friendly way of protecting citizens against maladministration, which has been deemed necessary because of the increasing impact of governments on citizens' lives. The ombudsman was thus seen as an antidote to “big government.” Ombudsman institutions had similar origins in former colonial states that inherited democratic institutions, such as Trinidad and Tobago, and Jamaica.

But as “ombudsmania” has grown, not surprisingly, so has its diversity (de Asper y Valdes 1999: 227). In the second half of the twentieth century, the institution rapidly spread around the globe. After the democratic revolutions in Spain and Portugal, a new wave of democratization dramatically lowered the number of authoritarian regimes. Structural democratic change (parliament, elections, etc.) occurred quickly, but citizens and public officials often were unaware of the basics of the rule of law, governmental integrity, and human rights; this need prompted an expansion of the role of ombudsmen in these new democracies. As Gregory puts it: “Underlying the widespread popularity of the institution, it seems, is the perception that the ombudsman concept is an idea whose time has come, and indeed come twice, with the second wave of Ombudsman Offices satisfying the need not so much for accountable administration in an age of ‘big government,’ its primary function in the West, but providing rather a valued weapon in the struggle against the recurrence of ‘bad governance’ in states ‘en route’ to liberal democracy from a totalitarian or authoritarian past” (Gregory 2002: 13).

The job titles assigned to those performing the ombudsman function illustrate this new role. In Spain, the constitution foresees a “Defensor del Pueblo,” a defender of the people. In Ukraine, there is a “Commissioner for Human Rights.” In Bulgaria, the annual report of the ombudsman must contain information on the respect of human rights. In Bosnia–Herzegovina and Albania, there is a “Human Rights Ombudsman.” However, most “new” ombudsmen also deal with everyday cases of maladministration, while some classical ombudsmen deal with “hard” human rights as well. It is therefore difficult to draw a sharp distinction between different sorts of ombudsmen.

In recent years, there has been a shift of private sector management techniques to the public sector (e.g., accrual accounting and performance management). The ombudsman concept, however, has been transferred from the public sector to organizations in the private and nonprofit sector, such as universities, the media, corporations, and hospitals.

With so many variations of ombudsman, it is not surprising that some academics wonder whether the concept has become “diluted” (Stieber 2000). Some public sector ombudsmen have tried to stop other organizations from using the term. In New Zealand, for example, there is a restriction on the use of the word itself. Only the chief parliamentary ombudsman is allowed to grant permission to use the title. The main concern is that the organizational ombudsman cannot give the same guarantees about independence and impartiality to those with complaints that the parliamentary ombudsmen can.

The debate about what it takes to be a “real” ombudsman is beyond the scope of this article and may never be entirely settled. Even within the field of public sector ombudsmen, there is considerable variation. Historically, the public sector ombudsman investigates citizens' complaints about virtually every part of public administration. But over time, some countries have introduced ombudsmen who are responsible for specific topics (e.g., gender equality, children) or specific parts of the public sector (e.g., transport, pensions).

In this article, we focus on the case of Belgian ombudsmen in the public sector at various levels of government. Belgium should be seen as a high‐income, established democracy in Western Europe, which means that ombudsmen in Belgium would fall into the category of “classical” ombudsmen.¹

The existing literature on ombudsmen is mainly descriptive or theoretical, and some commonly held assumptions have not been tested. Our goal here is to empirically investigate the ombudsman's impact on the relationship between citizens and government. Based on three surveys, we analyze whether the ombudsman is able to restore trust in government, which was one of the most important reasons (in Belgium and elsewhere) for introducing ombudsmen. We also examine the “social profile” of ombudsman complainants. In the reports made by ombudsmen, there is often an implicit assumption that these documents represent “what goes wrong within public administration.” But this is not necessarily true, because it may be possible that only a specific sector of the population files a complaint, while others — for various reasons — do not, which means that the complaints heard by ombudsmen may not be as representative as has been assumed. Therefore, in our research, we compare the profile of ombudsman complainants to that of average citizens to see how representative of the general population they are.

Belgium has had a national ombudsman law only since 1995, although there had been legislative proposals to establish one dating back to the 1960s (Anciaux 1990). In early 1997, the first Belgian ombudsman took office. (In Belgium, there are two ombudsmen at the national level: one serving those who speak Dutch, and one serving those who speak French.) To fully understand the establishment of the ombudsman office in Belgium, one must have knowledge of the (perceived) crisis in the relationship between citizens and government that arose in Belgium in the 1990s. The assassination of André Cools, the country's leading socialist politician, in 1991 initiated a series of events that seriously damaged the reputation of the entire Belgian political–administrative system. The election of November 24, 1991 was a second moment of crisis. This election constituted the breakthrough of the extreme right “Vlaams Blok” (“Flemish Block”) party. Politicians, journalists, and political scientists were surprised by the number of votes obtained by this party and its rise in the 1990s. Another event was the Agusta scandal that erupted in 1993. Several leading members of the Socialist party were found guilty of corruption in a bribery scandal that led to the resignation of Willy Claes as secretary‐general of the North Atlantic Treaty Organization. The final event that shocked the nation was a pedophilia crisis in 1996. In August 1996, Marc Dutroux and his companions were arrested for abducting, raping, and killing several young girls. The subsequent public anger transformed gradually. At first, anger was mainly aimed at Dutroux himself, but later, the government authorities were targeted as well. The perceived malfunctioning of the judicial system resulted in a “White March” attended by three‐hundred‐thousand citizens.

By 1997 and 1999, Belgium had become the country with the lowest trust in government in the Eurobarometer surveys (covering all European Union member states). But the crises brought about reforms in the public sector. It is within this framework that the Belgian government has taken measures to strengthen trust in the public sector. On the political front, a “new political culture” had to replace old political habits. To address the relationship between citizens and public administration, a United Kingdom‐like “citizen's charter” was launched in 1992, with limited success (Van de Walle, Thijs, and Bouckaert 2005). In the year 2000, a global administrative reform, called the “Copernicus‐reform,” was launched. This effort sought to mark a shift in the relationship between public administration and citizens. The name of the project was symbolic — just as Nicolaus Copernicus claimed that the earth was rotating around the sun, the goal of this reform effort was to create a situation where public administration rotated around the Belgian citizen and not the other way around (Thijs and Van de Walle 2005).

The ombudsman law must be seen within this framework. The perceived rift between the citizens and their government made the ombudsman law a priority during this period. The draft law explicitly mentions the confidence gap between citizen and government as an important reason for the introduction of the new institution, and this remains an important motive. In the 2003 job announcement for the new Belgian ombudsman, the first criteria listed was “the ombudsman should contribute to bridging the gap between citizens and government” (Chambre Des Représentants de Belgique 2005).

The same link between the ombudsman's role and the restoration of trust has also been made in Africa, the United States, and the Netherlands (Sargant 1964; Unruh 1968). Lack of trust was one of the main reasons cited for the introduction of ombudsmen at universities in the United States (Stieber 2000). Donald Rowat (2004) argues that an ombudsman at federal level in Canada would reduce a democratic deficit.

Empirical research on ombudsmen is scarce. The existing literature is either juridical or descriptive. Often, ombudsman institutions are introduced on premises that are, at best, theoretically (and not empirically) investigated. Only a very small percentage of organizational ombudsmen have provided comprehensive evaluations of their own work (Cominelli 2005).

In this article, we focus on the empirical examination of the effect of the ombudsman's individual role. Our first research question deals with the role of the ombudsman in strengthening trust in government, an issue that, to our knowledge, has never been addressed in empirical research. As outlined above, a major reason for introducing ombudsmen in different countries was to strengthen or restore trust in institutions. Organizations hope that by introducing an ombudsman, individual grievances will be resolved. The result could be an apology, redress, or giving the citizen what he or she is entitled to. Ultimately, if the organization did not perform as it should have, the ombudsman's intervention could lead to a change in the organization's behavior. The theoretical assumption is that an improvement in the performance of a particular governmental entity (i.e., doing what the citizen expects) could lead to higher trust in government as a whole. Within the literature on this topic, this is referred to as the “micro‐performance hypothesis” (Van de Walle and Bouckaert 2003).

Linking performance to satisfaction and trust, however, is problematic both theoretically and empirically. Different hypotheses are possible. For instance, it could be that individuals make independent judgments about government as a whole, on the one hand, and about individual agencies, on the other. Moreover, some particular agencies seem to have a stronger impact on citizens' opinions of government than do others. Empirical studies support the claim that the relationship between performance and trust is not straightforward (Kampen, Maddens, and Vermunt 2003). In a meta‐analysis of customer surveys in the United States, Thomas Miller and Michelle Miller (1991) found that fire services always get better grades than road repair services. An implication of this finding is that even a fire department that performs poorly according to objective performance measures will still receive a higher subjective rating than an effectively performing highway department. Other research indicates that citizens' opinions of the way public administration works are at least partly influenced by stereotypes (Van de Walle 2004).

Based on all this, we should not expect the ombudsman's work to have a significant effect on complainants' trust in government or public administration. Furthermore, given the fact that the Belgian ombudsman's office only receives a few thousand complaints every year, the effect of the ombudsman on directly restoring trust within the entire population could be expected to be limited at best.

A related second research question that we wish to address is the social profile of complainants. In a democratic welfare state, reaching all types of citizens is an important governmental goal. This may be even more true for the ombudsman. One reason for introducing (classical) ombudsmen was to respond to the increasing impact of government on citizens' lives. Ombudsmanship offered a simpler alternative to judicial procedures that, because of the expenses and other demands of litigation, are often beyond the reach of most citizens, especially those from lower socioeconomic levels. If wealthier and/or more educated people are the main consumers of the ombudsman's services, then the institution would have failed to achieve its objective to serve people at all socioeconomic levels. Moreover, when reporting on problems within public administration, the ombudsman will not be able to provide a truly representative overview because he or she will fail to uncover problems that mainly affect poorer and less educated citizens.

Theoretically, three conditions need to be met before someone files a complaint with an ombudsman (Felstiner, Abel, and Sarat 1980). First, the event must be noticed (“naming”) — an unperceived injurious experience must be transformed into a perceived injurious experience. This is not always obvious in relatively complex bureaucracies. Second, this experience must be transformed into a grievance; that is, the injured party must blame some person or entity for causing this injurious experience (“blaming”). In the last phase, this grievance is voiced to the organization perceived to be responsible for the injurious experience (“claiming”). As William Felstiner and his colleagues (1980) explain, most injurious experiences do not reach the last phase.

Concerning the social profile of complainants, some empirical research has already been undertaken. Available studies, however, are usually publicized only in the language of the country concerned, not in international academic journals, which makes it difficult to find literature that gives a complete overview of this topic. Nan Van Zutphen (2002) gives a thorough review of research in the United States, Germany, Spain, the Netherlands, the United Kingdom, and Belgium.² With one exception,³ the general profile of the user of the ombudsman's services that emerges from survey literature is “a man or woman in middle age with a higher education and a reasonable income and enough bureaucratical competences. . . . Members of other social groups also file a complaint with the ombudsman, but are underrepresented” (Van Zutphen 2002; translation by authors). As we previously mentioned, individuals may fail to take advantage of an ombudsman's services because they are unaware that the institution even exists. The individual may name the problem and blame a person or organization that he believes is responsible, but if he does not know of the existence of an ombudsman, he will not reach the claiming phase.

Research has found that more than 90 percent of Dutch people, for example, know about the institution of the national ombudsman (Hertogh 2000). But this kind of research results is difficult to interpret because participants in survey research often tend to give answers that are socially desirable (in this case: say they know what ombudsmen are). Sandra Loois and Marjan Sieben (2003) tried to take this phenomenon into account in their study by asking the participants where they could go if they had a complaint about the functioning of public administration: less than one in five mentioned the Dutch National Ombudsman.

These findings are part of a larger research tradition on the “Matthew effect.”⁴ The term, coined by American sociologist Robert K. Merton (1968), refers to situations where those who possess certain qualities or possessions will receive more, while less gifted individuals will receive less. Merton used the term to refer to a particular situation in the academic world. He found that when a well‐known scientist published something with someone who is relatively unknown, the former will almost always get the credit, even when the unknown academic is the first author. In cases of multiple independent simultaneous discoveries made by academics of “distinctly different rank,” celebrated scientists will receive more recognition (Merton 1968). The theory of the Matthew effect has also been used to describe various other phenomena (Merton 1988). In the field of social security for instance, it refers to the situation in which the bulk of social benefits goes to people who do not really need them. We hypothesize a similar Matthew effect in our research; that is, the ombudsman primarily receives complaints from higher status (wealthier and better educated) people.

To answer our questions, we analyzed three surveys of ombudsman clients in Belgium at three government levels. First, in 2004, we surveyed by mail 295 Dutch speakers who had made complaints to the Belgium ombudsman that had been dealt with in 2001. We did not investigate requests for information and inadmissible complaints or those complainants who had been referred to another ombudsman's office because those kinds of inquiries are not the “core business” of the Belgian ombudsman. The response rate to our survey was 58 percent, which can be considered reasonable.

The second and third surveys that we analyzed were conducted by Kaat Van den Audenaerde (2003). She investigated clients of the ombudsman of the Flemish (regional) public administration and of the ombudsman for the city of Mechelen.⁵ Data were collected by telephone. The response rate was neither high nor low for the ombudsman of Mechelen (46.4 percent, with one hundred fifty respondents). Some privacy issues arose in the data collection phase with the Flemish ombudsman, so the data collection was halted after only 181 complainants (of a sample of 726) were surveyed.

In all the three surveys, data sampling was based on official ombudsman records. The sample was representative of the population of ombudsman clients. In all the three questionnaires, the two main topics were the respondents' satisfaction with the ombudsman and the social profile of complainants.

Satisfaction with the ombudsman was measured in two ways. First, we asked respondents whether they were satisfied with the process that the ombudsman used to deal with their complaint. Second, we wanted to know if they were satisfied with the result.

In all the three surveys, more than three quarters of the respondents were satisfied with the way their complaint was dealt with by the ombudsman. This result is comparable to results from studies of satisfaction with university ombudsmen (Harrison 2004) but indicates greater satisfaction than was found in a study of complainants who used the services of a local government ombudsman in the United Kingdom (Market and Opinion Research International [MORI] Social Research Institute 1999). We should interpret these results cautiously, however, because the research methodologies are dissimilar in all cases except for the ombudsman of Flanders and of Mechelen.

Satisfaction with the result obtained by the ombudsman was significantly lower, which is consistent with what the MORI research has shown in the United Kingdom (MORI Social Research Institute 1999). This may reflect complainants' hopes that the ombudsman would be more favorable toward them in judging the situation. In reality, the mere fact that an ombudsman investigates a complaint does not mean that the complaint itself is legitimate. Another issue is that addressing some problems requires substantial structural changes, such as altering existing law, for example. For some complainants then, outcomes are uncertain. When comparing the satisfaction levels of the complainants in the three different Belgian studies analyzed here, we see fairly similar results. But again, comparing these results is difficult because of methodological differences between the three studies (Table One).

The second variable that is relevant here is the impact of the ombudsman on trust in government or public administration. The Belgian ombudsman study investigated trust in “public administration” (administratie or administrative functions of government) while the other surveys posed a question concerning complainants' trust in “government” (overheid refers to both the political and administrative [public administration] sides of government), thus including a political component in the question. The impact of these terminological differences should not be overestimated for two reasons. First, it should be noted that it was the explicit goal of policy makers to strengthen both trust in government, as well as trust in public administration. Second, survey research has shown that people's trust in government and public administration is of a general character. “A high level of trust in one institution tends to extend to other institutions” (Christensen and Laegreid 2003). Citizens frequently do not distinguish between different branches of government.

If the ombudsman's activities do indeed promote trust in government, we would expect that people who are satisfied with the result of their complaint would report higher trust ratings. Figure One displays the mean trust rating for each indicated level of satisfaction.⁶ It suggests that the link between satisfaction with the ombudsman and trust is “one way only.” Trust ratings are very low for those who are very dissatisfied, but it does not seem to make a difference whether complainants are rather dissatisfied or very satisfied. A hypothesis could be that the ombudsman does not strengthen complainants' trust in public entities but that he or she can weaken it. Another hypothesis is that dissatisfied complainants had an a priori negative attitude toward public institutions. Present research, however, will not adequately answer that question.

Correlation analysis confirms what theoretical and empirical contributions on this issue have found. There is, at best, a weak link between satisfaction with the ombudsman and trust in government. The correlation between satisfaction with the result of the ombudsman's intervention and trust in institutions is low in all cases, ranging from 0.099 (Belgian ombudsman) to 0.146 (Flemish ombudsman) to 0.173 (ombudsman of Mechelen). All correlations were significant at the 0.05 level.

Multivariate analysis for the Belgian ombudsman also supports the weak link between satisfaction and trust. One of our two research questions was whether the ombudsman has a direct effect on trust in government and public administration. Our results indicate that this effect is modest at best and should not be overestimated. As the literature on trust indicates, there seems to be no “silver bullet”— certainly not the ombudsman — available to governments to strengthen citizen's trust in them.

We now consider the profile of those people who filed complaints with the Belgian and Flemish ombudsmen. Although governments (and ombudsmen) in a social welfare state try to reach out to all citizens, a review of the existing literature allows us to hypothesize a “Matthew effect” in the profile of complainants.

First, we investigated a number of demographic indicators such as gender and age. Our results showed that the percentage of men who complained was about twice as large as the percentage of women. This is a general phenomenon that has been found in other empirical research on complainants (Jacobs‐Wessels 1995). Data for the Flemish ombudsman, however, show that the situation might be somewhat more complex. Quite often, when a woman had a problem with public administration, it was her husband who eventually filed the complaint (Van den Audenaerde 2003). Naturally, this makes the interpretation of these results more difficult.

Age is another variable for which we found significant differences between those who complain to the ombudsman and the general population. Young people were severely underrepresented, as were old people. This was compensated by an overrepresentation of the group of forty‐ to sixty‐year‐olds. Results did not greatly vary between the Belgian and Flemish ombudsman and the ombudsman of Mechelen.

A third relevant variable is the level of education. Again, we compare ombudsman complainants to the general population (reference group: eighteen‐ to eighty‐year‐olds). A near‐majority of complainants had a bachelor's or master's degree. On the other hand, the number of citizens with a lower level of education⁷ is severely underrepresented in both the Belgian and Flemish ombudsman samples.⁸ Other research had already found that citizens with a lower education are less often aware of the existence of ombudsmen (Loois and Sieben 2003) (Figure Two).

Occupation was the fourth variable analyzed. There was no large over‐ or underrepresentation of students, housewives, the unemployed, retirees, or workers. But when we looked exclusively at employed complainants, we found noteworthy differences between complainants and the larger population. In the Belgian ombudsman survey, we found a significant underrepresentation of blue collar workers.⁹

The last variable that we analyzed was the complainants' level of interest in politics. Respondents were asked to indicate on a four‐point scale how interested they were in politics. More than 70 percent of the Belgian ombudsman clients said they were “very” or “quite” interested (unfortunately, there was no similar question in the Flemish ombudsman survey). In the European Social Survey (a biennial multicountry survey),¹⁰ only about half of the citizens reported political interest. This finding is also reflected in the use of media — complainants reported that they watch the news and read high‐quality newspapers more often than the general population. A tentative hypothesis might be that the complainants' greater level of political sophistication makes them feel more comfortable using the ombudsmen system. But it is also possible that complainants are more often aware of the existence of ombudsmen, thus increasing their chances of claiming.

Our data point to one clear conclusion: the average complainant is not typical of the average Belgian in several key categories. Ombudsman clients in Belgium have a specific profile: they are likely to be highly educated, middle‐aged, white‐collared, politically interested men. Our results indicate that Belgian ombudsmen have had less success reaching socially disadvantaged and politically unsophisticated people. In most countries, the ability to fulfill the three conditions for making a complaint that we outlined above (naming, blaming, and claiming) are not equally distributed across the population. The use of ombudsmen is therefore associated with the Matthew effect in which more powerful members of society are able to take better advantage of an institution designed to support people at all levels of society. Even though ombudsmen are meant to be easily accessible, in reality, not all social groups access the ombudsman in the same proportion.

The number of ombudsman institutions has been growing at a noteworthy rate for several decades. Outside government, ombudsman offices have been adopted by banks, universities, corporations, and hospitals. As in numerous other countries in the world, Belgium introduced (classical) ombudsmen at different levels of government in the 1990s with the goal of reaching citizens who had turned away from government and strengthening their trust.

In this study, we investigated empirically the impact of ombudsmen in these two areas. The results are somewhat disappointing. The ombudsman can only strengthen trust in government and public administration to a limited extent. Because organizations often create ombudsmanship programs with the goal of restoring confidence in an organization, this research finding has implications for all ombudsmen, both public and private. Furthermore, we found the profile of complainants to be skewed: our results indicate that people from lower social classes are less likely to use the ombudsman's services.

Our findings have some interesting implications. For example, the annual reports of ombudsmen may not contain a representative picture of the problems within public administration. At the Third International Ombudsman Conference in 1984, the Swedish ombudsman mentioned that in some inspections, he encountered poignant situations that begged for improvement but about which he had received no complaints (Jacobs‐Wessels 1995). The results of our analysis of the link between satisfaction with the ombudsman and increased trust in institutions suggest that the ombudsman's role in restoring confidence is limited. Relying on the ombudsman to improve the relationship between an organization and its clients may therefore not be enough. Trust depends on numerous conditions, and the ombudsman can only influence a few variables.

What would improve ombudsmen's abilities to nurture trust in institutions? In our opinion, policy makers should change how they think of ombudsmen. As we have demonstrated, it would be unwise to see ombudsmen in Belgium as merely an instrument for improving individual relations between citizens and government. While we may never get a full picture of what goes wrong within any administration, ombudsmen can be used as an “early warning system.” In our view, ombudsmen should more consistently take the role of “change agents” who propose organizational improvements from a citizen's perspective (Wagner 2000).

A groundbreaking study in this field has been performed by Marc Hertogh (2001). He compared the impact of administrative courts with the impact of the ombudsman in controlling administrative action in the Netherlands. While ombudsmen and administrative courts are comparable in the sense that they investigate citizens' complaints, their judicial power is vastly different. Administrative courts are more coercive, whereas ombudsmen have a more “cooperative” style. In the long run, he found, because of this cooperative style, the policy impact of the ombudsman is much larger. In conclusion, our results indicate that we should not exaggerate the individual role of ombudsmen in Belgium, and policy makers should focus more on the collective role of ombudsmen, which could, we believe, be much greater.

As in all studies, this research has limitations. First, all of the surveys analyzed were conducted in just one small country. Ombudsmen in Belgium operate within specific cultural and legal contexts. We might hypothesize that differences in context (e.g., the context in Western countries versus the Indian caste system; common law versus civil law) would result in differences in findings. Also, although there is a Matthew effect in Belgium, this may not necessarily be the case in all countries. Clearly, these findings should be extrapolated to other countries with great caution.

Second, ombudsmen and politicians do have the ability to do something about the Matthew effect. For example, in Sweden, ombudsmen target specific (disadvantaged) groups. There is an ombudsman for equal opportunities (Jämställdhetsombudsmannen), children (Barnombudsmannen), people with disabilities (Handikappombudsmannen), against discrimination on grounds of sexual orientation (Ombudsmannen mot diskriminering på grund av sexuell läggning), and ethnic discrimination (Ombudsmannen mot etnisk diskriminering). It seems likely that because of their high visibility, these ombudsmen would attract more disadvantaged complainants. To our knowledge, however, research comparing the effectiveness of this system to the more traditional ombudsman systems does not exist.¹¹ Finally, traditional ombudsmen could maximize their coverage in the (popular) media, thus attracting a more diverse range of complaints; such actions might minimize the Matthew effect.