Previous research on public support for ‘independent experts’

Most of the literature on citizens’ attitudes towards government by non-traditional actors has employed survey questions referring to ‘experts’ more generally. For example, the European and World Values Surveys asked whether ‘having experts, not government, make decisions according to what they think is best for the country’ would be a good or bad way of governing the country (EVS, 2020). Studies on stealth democratic attitudes, a concept originally proposed by Hibbing and Theiss-Morse (Reference Hibbing and Theiss-Morse2002), asked whether ‘our government would run better if decisions were left to non-elected independent experts rather than politicians or the people’ (Bengtsson and Mattila, Reference Bengtsson and Mattila2009: 1034; Coffé and Michels, Reference Coffé and Michels2014: 5; Font et al., Reference Font, Wojcieszak and Navarro2015: 159; Webb, Reference Webb2013: 752). More recently, studies on technocratic attitudes asked whether ‘the best political decisions are taken by experts who are not politicians’ and whether ‘the problems facing my country require experts to solve them’ (Bertsou and Caramani, Reference Bertsou and Caramani2020: 6; Fernández-Vázquez et al., Reference Fernández-Vázquez, Lavezzolo and Ramiro2023: 10). Using the label of ‘independent experts’, studies have demonstrated widespread support for increasing the role of such non-traditional profiles in government among European publics (Bertsou and Pastorella, Reference Bertsou and Pastorella2016; Chiru and Enyedi, Reference Chiru and Enyedi2021; Lavezzolo et al., Reference Lavezzolo, Ramiro and Fernández-Vázquez2021).

When using the label of ‘independent expert’, these studies implicitly refer to technocrats. Technocrats are defined as experts with a high level of specialization derived from educational or professional qualifications, who do not have party-political experience (Caramani, Reference Caramani2017; Lavezzolo, et al. Reference Lavezzolo, Ramiro and Fernández-Vázquez2021). Such technocrats have increasingly been appointed as ministers in European governments over the past two decades (Vittori et al., Reference Vittori, Pilet, Panel, Paulis and Rojon2021). However, the independent experts or technocrats in government encompass a wide variety of individuals, some of whom might be more appealing than others (Dommett and Pearce, Reference Dommett and Pearce2019).

Only a few studies have tried to go beyond the generic label of experts by including items on support for government by different kinds of non-traditional actors. VanderMolen’s (Reference VanderMolen2017) study in the USA found that citizens are most favourable to governments run by fellow citizens or independent experts and least favourable to government run by businesspeople or bureaucrats, with elected politicians ranked in between. Another study in the USA by Hibbing and colleagues (Reference Hibbing, Theiss-Morse, Hibbing and Fortunato2021) demonstrated that the desire to empower businesspeople, military generals, and religious leaders is conceptually different from the desire to empower scientists, medical doctors, and independent experts.

Building on the aforementioned studies, our approach is to compare citizens’ preferences for a realistic set of ‘political outsiders’ as government ministers, namely high-ranking civil servants, university professors, and business leaders. These three profiles are contrasted with two ‘insider’ profiles, namely members of parliament and former ministers. Our use of the outsider/insider terminology is in line with previous work distinguishing between ministers without a parliamentary career and those whose main credentials are linked to party organizations and parliamentary life (Blondel and Thiébault, Reference Blondel and Thiébault1991; Costa Pinto et al., Reference Costa Pinto, Cotta and de Almeida2018). Former ministers might be considered insiders as the vast majority of government ministers are selected from the ranks of party leadership. Furthermore, our results will show that citizens’ attitudes towards MPs and former ministers are very highly correlated. Unlike Camerlo and Pérez-Liñán (Reference Camerlo and Pérez-Liñán2015), who distinguish between non-partisans with and without field expertise (respectively referred to as technocrats and outsiders), we refer to all profiles with non-partisan occupational backgrounds as outsiders.

However, we acknowledge that the outsider/insider categories are not mutually exclusive per se. For example, even among civil servants, university professors and business executives one might distinguish between insiders with experience in parliamentary and governmental circles and outsiders without any prior political experience. Nevertheless, we believe that comparing preferences for ministers from different career backgrounds represents a significant step forward in understanding citizens’ preferences for alternative profiles in government (especially in comparison to previous research which collectively refers to them as experts). Future research might build on our work by investigating how the nature or extent of an outsider’s ties to political institutions influence support for such outsiders in executive government positions.

Vice versa, we also acknowledge that among the MPs and former ministers referred to as insiders, there may be some with experience outside of parliamentary and governmental circles. However, studies conducted in several European countries have shown that ministers with a parliamentary background are much less likely to have worked as civil servants, university professors, or business leaders than those from outside parliament (Blondel and Thiébault, Reference Blondel and Thiébault1991; Costa Pinto et al., Reference Costa Pinto, Cotta and de Almeida2018). Therefore, we are fairly confident that respondents would perceive MPs and former ministers as having less training in occupations outside of politics than the three outsider profiles.

Previous research on who supports independent experts in government

Many studies have shown that the demand for independent experts in government is more prevalent among those with negative attitudes towards politics (Font et al., Reference Font, Wojcieszak and Navarro2015; Bertsou and Pastorella, Reference Bertsou and Pastorella2016; Del Río et al., Reference Del Río, Navarro and Font2016; Bertsou and Caramani, Reference Bertsou and Caramani2020; Chiru and Enyedi, Reference Chiru and Enyedi2021). Bertsou and Caramani’s study (Reference Bertsou and Caramani2020) in nine European democracies identified a group of ‘technocratic’ citizens, who combine the demand for more experts with a strong elitist and anti-politics stance. Technocratic citizens are a) anti-political because they criticize the role of elected representatives as intermediaries; b) expertise-oriented because they emphasize objective evidence-based policies; and c) elitist because they seek to empower those with superior academic or professional qualifications. Building on Bertsou and Caramani’s work, we will argue that preferences for political outsiders are stronger among citizens who are less trusting of representative actors and institutions, demand more government by expertise and reject the idea of encouraging ordinary citizens to participate in government. We therefore use political distrust, expertise orientations, and participatory orientations as proxies for the three sub-dimensions of technocratic attitudes (anti-politics, expertise, and elitism).

The demand for political outsiders may be rooted in disappointment with the performance of political parties and elected politicians. Independent experts are appealing because they are capable of taking informed decisions without having to toe the party line or defend voters’ interests. Elected politicians, on the other hand, are more concerned with securing re-election than with providing responsible solutions (Lavezzolo et al., Reference Lavezzolo, Ramiro and Fernández-Vázquez2021). According to a pure technocratic vision of politics, for every problem there is a rational, objective solution. This vision challenges the prevailing representative model of democracy, whereby policies are the result of negotiation and compromise between political parties representing different viewpoints (Caramani, Reference Caramani2017). Literature on the supply side of politics has shown that politicians portray themselves as outsiders to attract disgruntled voters (Guasti and Buštíková, Reference Guasti2020) and that technocratic ministers are often appointed as a face-saving tactic in the aftermath of political scandals (Wratil and Pastorella, Reference Wratil and Pastorella2018). Therefore, we expect that the appeal of political outsiders over traditional political profiles as ministers in government is stronger among individuals who are less trusting of representative actors and institutions, namely parliament, political parties, and politicians (H1a).

However, it is unclear whether politically distrusting individuals are open to any kind of alternative to traditional actors or whether they find some outsiders more appealing than others. University professors and business executives might be perceived as more independent from politics than high-ranking civil servants, who as public agents are employed by state apparatus. Political appointments of high-ranking civil servants are common in many western and southern European countries and while formal appointments are frowned upon in northern European countries, informal ones continue to take place, for example, through the creation of advisory bodies (Van der Meer et al., Reference Van der Meer, Steen, Wille, Raadschelders, Toonen and Van der Meer2007). Governments and civil service systems in European democracies generally suffer from a negative image and the downward trend in public confidence towards government is paralleled by a downward trend in public confidence towards the civil service (Van der Meer et al., Reference Van der Meer, Steen, Wille, Raadschelders, Toonen and Van der Meer2007). Therefore, we expect that university professors and business leaders are more appealing to politically distrusting individuals than civil servants, who may be perceived as more connected to politics (H1b). However, we also expect that the role of political distrust in explaining preferences for civil servants varies across countries, depending on the extent to which government administration is independent from party politics (Page and Wright, Reference Page and Wright1999; Van der Meer et al., Reference Van der Meer, Steen, Wille, Raadschelders, Toonen and Van der Meer2007).

Another reason why citizens might prefer outsiders over traditional political actors as ministers is that they want more expertise in government, which is a core component of technocratic attitudes. As explained by Bertsou and Carmani (Reference Bertsou and Caramani2020), expertise is about possessing superior educational and professional qualifications, taking evidence-based decisions, and acting independently of special interests. However, previous research on technocratic attitudes has asked whether citizens would prefer expertise more generally, instead of investigating whether the non-partisan profiles most often appointed to executive government positions appeal to those with expertise orientations. Our initial expectation is that civil servants, university professors, and business leaders are all the more appealing to expertise-oriented individuals than politicians. First, because these profiles are more often tied to specific fields, while politicians must legislate on a variety of topics. Second, politics might also be perceived as less meritocratic than other professions, where leaders are primarily selected based on their ability to carry out specific tasks and duties as opposed to their relatability or likeability. Outsiders, such as Silvio Berlusconi, Andrej Babiš, and Donald Trump, attempted to convince voters that the skills they gained as business entrepreneurs would make them more capable of running the state than career politicians (Castaldo and Verzichelli, Reference Castaldo and Verzichelli2020; Guasti, Reference Guasti and Buštíková2020). In his campaigns, Berlusconi projected himself as a self-made entrepreneur, promising to use his private sector managerial skills to improve the efficiency of the state (Castaldo and Verzichelli, Reference Castaldo and Verzichelli2020: 489). Therefore, we expect that the appeal of political outsiders is stronger among individuals who are more expertise-oriented (H2a).

However, citizens who demand more expertise in government are probably most favourable towards putting university professors in charge. Survey instruments developed to capture pro-expertise orientations (see Bertsou and Caramani, Reference Bertsou and Caramani2020) emphasize academic qualifications, scientific reasoning, and independence, all of which are characteristics often attributed to university professors. While it might seem obvious that citizens who value such characteristics would prefer to be governed by university professors, it is questionable whether they would find some of the other outsiders appointed to executive positions more appealing than politicians. Civil servants and business leaders might be considered experts on the basis of technical skills acquired working in government administration or the private sector, but they do not necessarily stand out as having the qualities ascribed to experts in the technocracy literature (Ganuza and Font, Reference Ganuza and Font2020). Therefore, we expect that university professors are more appealing to expertise-oriented individuals than civil servants or business leaders (H2b).

Departing from the usual observation that citizens are increasingly dissatisfied with politics, the literature on process preferences has emphasized that citizens may have different conceptions of who should govern and how. While stealth democrats prefer delegating more power to political outsiders such as independent experts and business leaders, participatory democrats demand a greater role for ordinary citizens in policymaking through instruments such as referendums and deliberative mini-publics (Webb, Reference Webb2013; Coffé and Michels, Reference Coffé and Michels2014; Font et al., Reference Font, Wojcieszak and Navarro2015; Bengtsson and Christensen, Reference Bengtsson and Christensen2016; Del Río et al., Reference Del Río, Navarro and Font2016; Gherghina and Geissel, Reference Gherghina and Geissel2020). Stealth democratic and participatory orientations are assumed to be incompatible, as stealth democrats want politics to be different, but do not seek to participate themselves (Hibbing and Theiss-Morse, Reference Hibbing and Theiss-Morse2002). However, while previous research has emphasized theoretical incompatibility between people-centric and expert-centric models of democracy, few studies have compared what expertise-oriented and participatory-oriented citizens think about empowering specific kinds of political outsiders.

Participatory orientations might be negatively related to preferences for political outsiders for a few reasons. First, unlike politicians, civil servants, university professors, and business leaders do not derive their legitimacy from courting voters and are therefore impervious to popular demands. Second, by virtue of their specialized knowledge and expertise, these outsiders constitute a social and economic elite whose political views may differ from those of the majority (Fernández-Vázquez et al., Reference Fernández-Vázquez, Lavezzolo and Ramiro2023). Third, research on techno-populism during the COVID − 19 pandemic has shown how technocratic expertise is used to curb democratic freedoms. For example, in the Czech Republic, businessman-turned-prime-minister Andrej Babiš used the language of technocratic competence and public health to justify his decision to ban protests (Guasti, Reference Guasti and Buštíková2020). Therefore, we expect that the appeal of political outsiders over traditional political actors in government is weaker among participatory-oriented individuals who demand a greater role for ordinary citizens in policymaking (H3a).

However, participatory-oriented individuals may distinguish between different kinds of political outsiders. While this has not been tested in previous research, some studies suggest that participatory-oriented individuals and university professors share similar views. For example, individuals preferring a greater role for citizens in political decision-making are found to express greater tolerance towards minorities (Christensen and von Schoultz, Reference Christensen and Hesstvedt2019) and greater support for addressing economic inequalities (Rojon and Rijken, Reference Rojon and Rijken2020). Similarly, university professors are found to be more accepting of immigrants and more in favour of reducing income differences than other occupational groups, while the reverse is true for CEOs (van de Werfhorst, Reference Van de Werfhorst2020). Participatory-oriented individuals might therefore perceive university professors as upholding their commitment to equality and diversity. Furthermore, participatory-oriented individuals might expect university professors to support their demand for direct citizen participation, as universities are generally regarded as incubators of civic participation and intellectuals are historically known for their role in civic movements (Dodson, Reference Dodson, Gross and Simmons2014). Therefore, we expect that participatory-oriented individuals are less negative about university professors as ministers than civil servants or business executives (H3b).

On the other hand, it could also be argued that participatory-oriented individuals are more favourable towards business leaders than university professors or civil servants as ministers. Business leaders might be perceived as more external and less instrumental to established political parties than the other two profiles. First, because business leaders are the least common type of non-partisan minister in European cabinets among the three (Vittori et al., Reference Vittori, Pilet, Panel, Paulis and Rojon2021). Second, because political parties have regularly co-opted civil servants and university professors to improve their image and promote their policy reforms (Boswell, Reference Boswell2008; Helms, Reference Helms2022; Hesstvedt and Christiansen, Reference Hesstvedt and Christiansen2022). By contrast, prominent examples of business leaders entering politics, such as Trump, Berlusconi, Babiš, or Stronach, have showcased their potential for challenging established political parties. Therefore, we expect that participatory-oriented individuals are less negative about business leaders as ministers than university professors or civil servants (H3c).

Finally, while the literature has pointed to groups of individuals emphasizing different visions of democracy, the common denominator is that these individuals are all disillusioned with politics as usual. This means, first and foremost, that expertise and participatory-oriented citizens who are very trusting of representative actors and institutions are probably not as keen on replacing politicians with outsiders as those who are very distrusting. Second, this means that although expertise and participatory-oriented citizens emphasize different ways of governing, the more distrusting they are, the more they will prefer any alternative to run-of-the-mill politicians. In other words, political distrust acts as a moderator of the relationship between specific visions of democracy and preferences for political outsiders. Therefore, we expect that political distrust increases preferences for all three political outsiders among both expertise-oriented (H4a) and participatory-oriented (H4b) individuals.

Data

The data were collected through a CAWI survey fielded in 15 European countries during winter 2021–2022. On one hand, the context of the COVID-19 Pandemic may have encouraged more favourable attitudes towards experts (as suggested by Lavezzolo et al., Reference Lavezzolo, Ramiro and Fernández-Vázquez2021). On the other hand, the survey was administered after most of the pandemic lockdowns and vaccination rollouts in Europe, meaning epidemiologists were no longer in the spotlight. The countries were selected with the objective of variation in regions, political systems, and levels of experience with technocratic ministers in government (based on Vittori and colleagues’ Reference Vittori, Pilet, Panel, Paulis and Rojon2021 Technocratic Ministers Data set). A quota sampling strategy ensured sufficient representation of persons from different socio-demographic groups based on four key characteristics: age, sex, education, and region of residence. Each country sample (around 1,550 respondents) was weighted to match the distributions on these socio-demographic characteristics in the general population. The total sample includes around 24,000 respondents, after excluding trackers, speeders, and inattentive respondents who accounted for less than 5% of the raw data.Footnote ¹

Dependent variables: preferences for political outsiders in government

Although most studies measure preferences for political outsiders in government (rather than traditional politicians) by asking whether decisions should be taken by ‘independent experts’ as opposed to politicians, this label might encompass a variety of actors (Font et al., Reference Font, Wojcieszak and Navarro2015). We focus, instead, on capturing preferences for ministers from different non-political backgrounds that are commonly represented in European governments. A unique inventory of non-elected, non-partisan ministers in European democracies since 2000 (Vittori et al., Reference Vittori, Pilet, Panel, Paulis and Rojon2021) demonstrated that high-ranking civil servants, university professors, and business executives are the most common outsiders, respectively, accounting for 32, 20, and 15% of non-partisan ministers in Europe (see Figure 1). Therefore, we compare support for these three profiles as ministers with support for two traditional political profiles, namely MPs and former ministers.

Figure 1. Distribution of non-elected, non-partisan ministers in European democracies between 2000 and 2020 by occupational background (Vittori et al., Reference Vittori, Pilet, Panel, Paulis and Rojon2021).

In our original survey questions, respondents were asked how they would feel about each of the three non-political profiles as well as the two political profiles, becoming ministers in their country’s government, on a scale ranging from very negative (0) to very positive (10). By presenting respondents with five independent items, we allow them to give similar ratings to two or more profiles should they perceive them as equally (un)suitable for ministerial positions. As shown in Table 1, support for the two political profiles, MPs and former ministers, is highly correlated, which is a strong reason for combining them into a scale averaging respondents’ scores on the two items. A strong correlation between MPs and former ministers was observed in all fifteen countries (see online Appendix I).

Table 1. Correlations between each of the profile ratings in the pooled sample (N = 24,000)

The correlations in Table 1 show that university professors and business leaders are rated quite differently to MPs and former ministers, while high-ranking civil servants are rated somewhat similarly to the two political insiders. The slightly stronger correlations between MPs/former ministers and civil servants, relative to the other outsiders, might be explained by cross-country variation in the extent to which the civil service is independent from party politics. This is supported by the finding that civil servants were rated less similarly to MPs and former ministers in countries that are known for having less politicized bureaucracies, such as Denmark, Finland, Ireland, and the UK (see online Appendix II). However, as civil servants are traditionally expected to uphold neutrality (Hojnacki, Reference Hojnacki, Bekke, Toonen and Perry1996) and the civil service is the most common career path for non-elected ministers in Europe (Vittori et al., Reference Vittori, Pilet, Panel, Paulis and Rojon2021), we decided against combining it with the two traditional political profiles.

Our original items on feelings about civil servants, professors, and business leaders becoming ministers do not capture (dis)preferences for political outsiders, as they do not show whether respondents rated these profiles more or less favourably than the political ones. Therefore, we created three dependent variables based on the deltas between respondents’ ratings of politicians (average score for MPs and former ministers) and their ratings of civil servants, professors, and business leaders, respectively. We argue that examining differences between scores attributed to the five profiles makes sense as respondents were asked to rate the five profiles successively. Therefore, if a respondent gave a score of 4 to the first profile and a score of 7 to the second one, they were deliberately presenting themselves as more positive about the latter profile than the former.

The deltas were then recoded as categorical variables differentiating between 0) respondents who rated politicians at least one point (out of 10) higher than the outsider; (1) respondents who rated the outsider at least one point (out of 10) higher than politicians; and (2) respondents who rated both profiles similarly (less than one point difference between the two). Using deltas as opposed to raw scores means that respondents who rated political outsiders below the mid-point of the scale (but still higher than politicians) might be labelled as preferring outsiders. However, a cross-tabulation of our dependent variables with the original survey items demonstrated that only between 10 and 16% of respondents labelled as preferring a civil servant, professor, or business leader actually rated that same profile below the mid-point of the scale (5).

Independent variables

The hypotheses were tested with three key predictors: political distrust, expertise orientations, and participatory orientations. Political distrust is a scale averaging respondents’ feelings of distrust (1 = high trust/5 = no trust) in the three key organs of representative democracy: parliament, political parties, and politicians. These items formed a reliable scale in all countries, with alphas ranging between 0.8 and 0.9.

Expertise orientations is a scale averaging respondents’ agreement (1 = strongly disagree/5 = strongly agree) with four statements from Bertsou and Caramani’s (Reference Bertsou and Caramani2020) expertise dimension of technocratic attitudes: ‘the leaders of my country should be more educated and skilled than ordinary citizens’; ‘social problems should be addressed based on scientific evidence, not ideological preferences’; ‘the problems facing my country require experts to solve them’; and ‘our country would be better governed if important decisions were left to independent experts’. These items formed a reliable scale in all countries with alphas ranging between 0.6 and 0.8.

Participation orientations is a scale averaging respondents’ agreement (1 = strongly disagree/5 = strongly agree) with three statements about citizens being able to take political decisions either through referendums, deliberative mini-publics, or more generally (see Appendix III for full wording). These items formed a reliable scale in all countries with alphas ranging between 0.6 and 0.8.

Several variables that might influence both our dependent and independent variables were included as controls, namely age, sex, education, feelings about one’s household income, political interest, and left-right economic and cultural value orientations. Several studies demonstrated that expert preferences are stronger among younger, higher educated, and wealthier individuals who might feel that their views are more in line with those of experts (Webb, Reference Webb2013; Bertsou and Pastorella, Reference Bertsou and Pastorella2016; Rapeli, Reference Rapeli2016; Fernández-Martínez & Font-Fábregas, Reference Fernández-Martínez and Font-Fábregas2018; Berstou and Caramani, Reference Bertsou and Caramani2020; Chiru and Enyedi, Reference Chiru and Enyedi2021). Whereas political interest is negatively correlated with stealth democratic attitudes (Bengtsson and Mattila, Reference Bengtsson and Mattila2009; Webb, Reference Webb2013), it is positively correlated with technocratic attitudes (Bertsou and Caramani, Reference Bertsou and Caramani2020).

Although technocracy denies the relevance of ideological debates and technocrats are idealized as ‘impartial’ actors (Putnam, Reference Putnam1977; Centeno, Reference Centeno1993), there is a historical overlap between technocratic cabinets and neoliberal reforms (Centeno and Silva, Reference Centeno and Silva1998: 4–6). This might explain why support for ‘non-elected experts’ is often found to be stronger among individuals identifying as right wing (Bertsou and Pastorella, Reference Bertsou and Pastorella2016; Del Río et al., 2016; Bertsou and Caramani, Reference Bertsou and Caramani2020; Lavezzolo et al., Reference Lavezzolo, Ramiro and Fernández-Vázquez2020; Chiru and Enyedi, Reference Chiru and Enyedi2021). However, because the policy space in Europe is structured by an economic and cultural dimension, the issues citizens associate with left and right may differ (Kitschelt, Reference Kitschelt2004). Therefore, the relationship between left-right value orientations and political outsiders might depend on a) which dimension citizens have in mind and b) who the outsiders are. For this reason, we include attitudes towards income redistribution and immigration as proxies for the economic and cultural dimensions, with higher values indicating more left-wing positions, as opposed to using the standard left-right self-placement scale.Footnote ²

Method of analysis

First, we begin by comparing preferences for the three political outsiders as ministers across the 15 countries, based on the distribution of respondents by preference group: prefers one of the outsiders, prefers politicians, prefers neither.

We then turn to comparing the characteristics of citizens preferring each political outsider, drawing from the literature on technocratic attitudes, which connects the demand for independent experts to anti-politics, expertise, and elitism (Bertsou and Caramani, Reference Bertsou and Caramani2020; Fernández-Vázquez et al., Reference Fernández-Vázquez, Lavezzolo and Ramiro2023). This is done by modelling the effects of political distrust, expertise orientations, and participatory orientations (as proxies for the three dimensions of technocratic attitudes) on membership to one of the preference groups using three separate multinomial logistic (MNL) regressions. In our analyses, we compare the effects of our predictors on the probability of preferring each political outsider versus preferring a politician, while excluding respondents who showed no preference (see Appendix IV for the full model specification including the neutral groups). The analyses presented are based on the pooled sample combining data from all 15 countries, with country-fixed effects to control for potential heterogeneity. However, we also ran the models separately for each country to check the robustness of our key predictors (see Appendix VI).

Finally, to investigate whether the effects of expertise and participatory orientations on preferences for outsiders are moderated by political distrust (H4a & H4b), we include interactions between each of these predictors and political distrust in a logistic regression analysis of the probability of preferring outsiders as opposed to politicians. Respondents with no preference are also excluded from the latter analyses as we did not formulate expectations for how they differ from those preferring politicians.

Descriptive results

The stacked bar charts in Figure 2 compare the distribution of respondents by preference group (prefers outsiders/no preference/prefers politicians) between countries.

Figure 2. Distributions of preferences for political outsiders by country.

Our first observation is that citizens in all 15 countries express much stronger preferences for academics over politicians than for business leaders or high-ranking civil servants. Those preferring professors outnumber those preferring politicians in 14 out of 15 countries, while those preferring civil servants or businesspeople are only more prevalent in a few countries, namely Poland, Italy, Spain, Ireland, Belgium, and Greece. Citizens from Poland, Italy, Spain, and Ireland generally prefer all three outsiders, while citizens in the remaining countries prefer specific kinds of outsiders, e.g., academics.

Our second observation is that the group of citizens expressing no preference for either profile is much larger when it comes to civil servants. The finding that civil servants are more often rated similarly to politicians than academics or business leaders, goes in the direction of our claim for H1b, that the latter two profiles are perceived as more distant from politicians than civil servants.

Country differences in citizens’ preferences for political outsiders might be explained by several country-level factors, the most obvious of which is the prevalence of political outsiders in previous governments. Therefore, for each country, we plotted the percentage of non-partisan ministers with backgrounds as civil servants, professors, and business executives for the period 2000–2021 against the percentage of respondents preferring each political outsider over a politician as minister. The three scatterplots in Figure 3 seem to suggest a slightly positive relationship, with the exception of a few outliers. For example, citizens from the Czech Republic and Bulgaria are comparatively less enthusiastic about outsiders despite having been governed by technocratic ministers more often than citizens from other European countries. Research from the Czech Republic and Bulgaria suggests that technocratic ministers struggle to present themselves as truly independent from party politics (Hanley, Reference Hanley2018). Disillusionment with outsiders might also play a role, for example, citizens from the Czech Republic, where businessman-turned-prime-minister Andrej Babiš was recently accused of corruption and fraud, are the least favourable towards business leaders as ministers (see Figure 2). Therefore, while our results suggest a potential ‘technocracy addiction effect’, the inability of political outsiders to maintain an image of disinterestedness and impartiality might contribute to certain outliers.

Figure 3. Percentage of non-partisan, non-elected ministers since 2000 (Vittori et al., Reference Vittori, Pilet, Panel, Paulis and Rojon2021) by percentage of respondents preferring outsiders.

Another potential explanation for country differences in preferences for political outsiders might be economic and political crises. Research on the emergence of technocratic cabinets and ministers shows that they often materialize as a result of party failure, such as, for example, when governing parties are unable to form a cabinet or prevent imminent economic downturns (Brunclík, Reference Brunclík2015; Costa Pinto et al., Reference Costa Pinto, Cotta and de Almeida2018; Alexiadou and Gunaydin, Reference Alexiadou and Gunaydin2019). This would explain why citizens preferring political outsiders are much more prevalent in countries like Italy, Spain, and Greece but also in Ireland, Belgium, and Austria.

A last country-level factor that might explain the differences we observe in preferences for civil servants across countries is the degree of independence of the civil service. As elaborated above, in some of the countries covered, the civil service is very independent and partisan ties do not play a role in the recruitment of civil servants. In other countries, government bureaucracy is much more politicized (see Page and Wright, Reference Page and Wright1999; Van der Meer et al., Reference Van der Meer, Steen, Wille, Raadschelders, Toonen and Van der Meer2007). At first glance, the descriptive results in Figure 2 suggest that preferences for civil servants over politicians as ministers are lower, rather than higher, in countries that are known for having a more independent government bureaucracy (i.e., Denmark, Netherlands, and Finland). By contrast, the share of respondents preferring civil servants over politicians is much greater in countries where ministerial cabinets appointed by the ruling parties are relatively common (i.e., Italy and Spain) (Van der Meer et al., Reference Van der Meer, Steen, Wille, Raadschelders, Toonen and Van der Meer2007).

While the descriptive results presented above shed light on cross-country differences in citizens’ preferences for political outsiders, systematically testing the role of country-level factors would require a greater number of countries. Furthermore, we found that individual-level factors play a much greater role in explaining citizens’ preferences for political outsiders. Given that the MNL regression estimates, presented in the next section, do not provide a reliable measure of explained variance (only the pseudo-R-squared), we also included the OLS regression estimates of citizens’ preferences for political outsiders in Appendix VII (based on continuous dependent variables representing the difference between outsiders and politicians prior to recoding into preference groups). Models including only country dummies explain between 3 and 6% of the variance in preferences for the three non-partisan ministers, while models including both country dummies and individual-level predictors explain between 6 and 18% of the variance in preferences for non-partisan ministers. Therefore, the socio-demographic and attitudinal differences between citizens, elaborated in the next section, appear to offer more compelling explanations for why citizens prefer non-partisans as ministers. It is worth noting, however, that the R-squared is generally lower for models estimating preferences for civil servants, suggesting that our predictors are less useful for explaining why citizens preferred this profile over politicians.

Results: who prefers which political outsiders as ministers

Table 2 presents the MNL regression estimates of preferences for civil servants, university professors, and business leaders over politicians as ministers in Models 1, 2, and 3, respectively. The full model specification, including the effects of our control variables and the effects of our predictors on the probability of showing no preference for either profile, is provided in Appendix IV. The relative risk ratios (RRR) in Models 1 to 3 show how the risk of preferring a political outsider over a politician as minister changes with the predictors in question. A RRR greater than 1 indicates that the risk of preferring an outsider increases as the predictor increases, while a RRR less than 1 indicates that the risk of preferring an outsider decreases as the predictor increases.

Table 2. Multinomial logistic regression estimates of preferences for political outsiders as ministers (N = 24,000)

Note: The samples are weighted to match the distributions on age, sex and education in the general population and all models include control variables and country-fixed effects.

***P < 0.001, **P < 0.01, *P < 0.05.

The results confirm that preferences for political outsiders over politicians as ministers are stronger among individuals who are more distrusting of representative institutions, such as parliament, political parties, and politicians (H1a). This effect is significant for all three outsiders but even stronger for university professors and business executives, providing support for our claim that high-ranking civil servants are not as appealing to politically distrusting individuals as the other two outsiders (H1b). A one unit increase in distrust is associated with a 60–70% increase in the probability of preferring professors or business executives over politicians compared to a 38% increase in the probability preferring civil servants. The coefficient plot in Figure 4 provides a visual representation of the strength of our key predictors on the probability of preferring each political outsider over politicians, based on three separate logistic regressions (coefficients provided in Appendix X). By-country analyses in Appendix VI demonstrate that distrust fails to predict preferences for civil servants in France, Germany, and The Netherlands, where respondents also rated civil servants more similarly to MPs and former ministers than in other countries (see Appendix II). By contrast, distrust increases preferences for professors and business leaders in all countries except Bulgaria, where politically distrusting individuals do not prefer these outsiders over politicians, even though they have often held ministerial roles (see Figure 3). The chi-squared tests for the by-country analyses in Appendix VI show that our independent variables have a lower predictive in Bulgaria and Poland compared to other countries. Therefore, preferences for political outsiders in Bulgaria and Poland may be driven by other attitudes not accounted for in this study.

Figure 4. Coefficient plot for the logistic regression estimates of preferring each political outsider over politicians.

The results also confirm that preferences for political outsiders as ministers are stronger among individuals who are more positively oriented towards the use of expertise in government. Expertise orientations are significantly related to all three profiles (H2a), but even more strongly related to preferences for university professors (H2b). A one unit increase in expertise orientations is associated with a 43% increase in the probability of preferring professors over politicians but only a 12 and 9% increase in the probability of preferring civil servants and business leaders, respectively. By-country analyses in Appendix VI demonstrate that expertise orientations only increase preferences for civil servants and business leaders in a few countries. By contrast, being more positively oriented towards the use of expertise in government leads to preferences for university professors in 10 out of 15 countries.

While using a scale including items measuring support for experts to explain preferences for civil servants, professors, and business leaders might seem like an endogenous strategy, the step-wise regressions in Appendix IV show that excluding expertise orientations from the models does not affect our findings (the coefficients of our other predictors change very little). Furthermore, the Bertsou and Caramani scale also includes items emphasizing scientific reasoning, academic credentials, and independence, which we show are more strongly connected to preferences for university professors as ministers. It is possible that civil servants and business leaders are associated with other qualities that are not captured by the conceptualization of expertise orientations in the technocracy literature.

Contrary to expectations (H3a), preferences for political outsiders over politicians as ministers are not weaker but stronger among individuals who are more positively oriented towards involving citizens in political decision-making. This challenges the claim that citizens who are favourable towards empowering outsiders such as civil servants, professors, and business leaders are adverse towards people sovereignty (Bertsou and Caramani, Reference Bertsou and Caramani2020). A potential explanation for this is that participatory-oriented individuals perceive the appointment of non-traditional actors as a way of broadening participation in government, or of challenging the monopoly of elected politicians over policy making. In any case, it confirms the need for further empirical research on the kinds of profiles participatory-oriented individuals would like to see in government.

Finally, we observe that participatory orientations have the strongest association with preferring business leaders as ministers (in line with H3c) and the weakest association with preferring university professors (contrary to H3b). A one unit increase in participatory orientations is associated with a 52% increase in the probability of preferring business leaders over politicians compared to a 35 and 25% increase in the probability of preferring civil servants and professors, respectively. Claiming that participatory-oriented individuals prefer business leaders over the other outsiders based on a slightly larger coefficient might be misleading. This is because we do not know whether people who prefer business leaders differ significantly on participatory orientations from those who prefer civil servants over university professors. Therefore, as a robustness check, we tested the effects of participatory orientations on preferring business leaders over university professors (Appendix VIII) and over civil servants (Appendix IX). This was done by creating separate dependent variables capturing the difference between respondents’ ratings of business leaders and their ratings of the other two outsiders. The results show that participatory-oriented individuals are significantly more likely to express a preference for business leaders over the other two non-partisans as ministers, confirming H3c.

By-country analyses in Appendix VI demonstrate that participatory orientations increase preferences for business leaders in most countries but only increase preferences for the other two outsiders in half of the countries (e.g., BE, DK, FR, DE, IE, IT, NL, and UK). Interestingly, participatory-oriented individuals from Bulgaria and Poland, where non-partisan ministers are more prevalent than in other countries (see Figure 3), do not find any of the three outsiders more appealing than politicians. Contrary to expectations, these findings indicate that being in favour of greater participation is complementary to, rather than incompatible with, preferring non-partisan ministers. Furthermore, they suggest that profiles that are regularly co-opted by political parties, namely university professors and business leaders, are slightly less appealing to participatory-oriented individuals than business leaders.

Results: political distrust as a moderator of expertise and participatory orientations

We now turn to testing whether the effects of expertise and participatory orientations on preferences for political outsiders are moderated by political distrust (H4 a & b). The interaction effects are more easily interpretable through the contour plots in Figure 5 than through the regression estimates, provided in Appendix X. Using the margins command in Stata, we calculated the predicted probability of preferring each outsider over politicians for all combinations of expertise orientations and distrust (Figure 5, left-hand plots) and all combinations of participatory orientations and distrust (Figure 5, right-hand plots). Darker shades demonstrate higher probability of preferring an outsider over politicians. Curved lines demonstrate that the effects of expertise and participatory orientations differ across levels of political distrust, thereby suggesting a significant interaction effect.

Figure 5. Predicted probabilities of preferring outsiders by distrust and expertise orientations.

Note: the shades do not correspond to the same predicted probabilities across graphs as the effects differ in size.

Starting with the expertise by distrust interactions, the left-hand plots show that someone who is very trusting (<2 on distrust) and very expertise-oriented (>4 on expertise) has an almost 30% chance of preferring civil servants over politicians, while someone who is equally trusting but rejects expertise (<2 on expertise) has a 10% chance of preferring civil servants. Similarly, a very trusting pro-expertise individual has a 40% chance of preferring professors, while a very trusting anti-expertise individual has a 20% chance of preferring professors. This demonstrates that expertise orientations lead to preferences for civil servants and professors as ministers even among those who are confident in representative actors and institutions. By contrast, expertise orientations do not increase preferences for business leaders among politically trusting individuals.

The curvature of the lines in the left-hand plots for civil servants and professors suggests, in line with expectations (H4a), that political distrust boosts preferences for these outsiders among expertise-oriented individuals. Someone who scores high on both expertise orientations (>4 on expertise) and political distrust (>4 on distrust) has a 20% greater chance of preferring civil servants and a 40% greater chance of preferring professors than someone who scores equally high on expertise orientations but very low on distrust (<2 on distrust). The straight line in the left-hand plot for business leaders suggests, contrary to expectations (H4a), that the effect of expertise orientations on preferences for this outsider remains relatively stable across different levels of distrust. Indeed, the regression estimates (Appendix X) demonstrated that the expertise by distrust interaction only predicts preferences for civil servants and professors.

Turning to the participatory by distrust interactions, the right-hand plots show that someone who is very trusting (<2 on distrust) and very participatory-oriented (>4 on participation) has an almost 40% chance of preferring both civil servants and business leaders over politicians, while someone who is equally trusting but rejects participation (<2 on participation) has a less than 20% chance of preferring these outsiders. This demonstrates that supporting an expanded role for citizens in political decision-making leads to preferences for civil servants and business leaders even among those who are confident in representative actors and institutions. By contrast, participatory orientations do not increase preferences for professors among politically trusting individuals.

The curvature of the lines in the right-hand plots for civil servants and business leaders suggests, in line with expectations (H4b), that political distrust boosts preferences for these outsiders among participatory-oriented individuals. Someone who scores high on both participatory orientations (>4 on participation) and political distrust (>4 on distrust) has an almost 20% greater chance of preferring civil servants and an almost 40% greater chance of preferring business leaders than someone who scores equally high on participatory orientations but very low on distrust (<2 on distrust). The relatively straight lines in the right-hand plot for professors suggests, contrary to expectations (H4b), that the effect of participatory orientations on preferences for this outsider remains relatively stable across different levels of distrust. Indeed, the regression estimates (Appendix X) demonstrated that the participatory by distrust interaction only predicts preferences for civil servants and business leaders.

Therefore, our results show that expertise-oriented individuals are even more likely to prefer civil servants and professors over politicians as ministers when they are distrusting of representative actors and institutions. Furthermore, participatory-oriented individuals are even more likely to prefer civil servants and business leaders when they are distrusting of representative actors and institutions. However, political distrust is not a prerequisite; politically trusting expertise-oriented individuals also express preferences for civil servants and professors, while politically trusting participatory-oriented individuals also express preferences for civil servants and business leaders.