The Good Politician and Political Trust: An Authenticity Gap in British Politics?

There are three broad sets of qualities that citizens might expect politicians to display: competence, integrity and authenticity. To be authentic, a politician must be judged to be in touch with the lives and outlooks of ordinary people and previous research has suggested that this expectation has grown more prevalent in recent times. In this article, we use survey evidence from Britain – from citizens, parliamentarians and journalists – to explore which groups are prone to judge politicians by which criteria. While all groups give the highest absolute importance to integrity traits, we establish that distrusting citizens are significantly more likely to prioritise authenticity. For political elites and journalists, we find indications that authenticity is less valued than among citizens: politicians place more relative importance on integrity traits while journalists value competence most. We reflect on these findings and how they help us understand the growing crisis afflicting British politics.


Introduction
The personality and qualities of political leaders has long been acknowledged as an important element of politics and influence on political attitudes and behaviour (Declercq et al., 1975;Laswell, 1930;Regenstrei, 1965). In more recent decades, academic attention has turned towards how the expectations that citizens have of politicians determine these important dynamics (Garzia, 2011;Pancer et al., 1999). This increasing attention has been accompanied by the argument that these expectations and evaluations have become even more important over time, as partisanship in the electorate has declined dramatically (Dalton and Wattenberg, 2000;Mair and Van Biezen, 2001). Citizens are less tied by loyalty to a party or candidate and thus might be more engaged in a process of judgement about who to lend their support to at each election (Dalton, 1984(Dalton, , 2009). The dynamics of interaction between politicians and citizens have also changed as media technology develops rapidly and party organisations shift, fundamentally reshaping our political landscapes (Garzia, 2011;McAllister, 2007;Manin, 1997). These new spaces for interaction can support a greater focus by citizens on the personal qualities of the leader.
In this study, we explore the structure of citizens' expectations towards politicians' personality traits and the relationship between these expectations and political trust, using a 2018 representative sample of British adults. We compare these with the results of a small-N survey conducted among political elites in the UK in 2017 to examine whether expectations are consistent between citizens and elites. We open the article with a focus on three traits: competence, integrity and authenticity. The first two categories are widely understood and used. The same cannot be said for the third. We clarify the scope of authenticity and note the evidence that it may be becoming an increasingly important criterion by which citizens judge politicians. In the second section, we explore the potential connections between political trust and authenticity. After outlining our research strategy, we present our findings. The concluding discussion explores the implications of our findings given the extensive lack of political trust and confidence that has characterised the attitudes of British citizens for at least the past decade (Clarke et al., 2018;Stoker, 2017;Whiteley et al., 2016).

The Qualities of Political Leaders: Competence, Integrity and Authenticity
There has been some variability in the findings and terminology of previous studies on the personal traits of politicians: while most of these find two distinct dimensions relating to competence on one hand and integrity on the other, there is less consistency on other potential dimensions (Brown et al., 1998;Garzia, 2011;Miller et al., 1986). A third dimension has been called 'charisma', described as the ability of leaders to persuade voters, but also sometimes including traits such as warmth and humility (Miller et al., 1986;Seijts et al., 2015). Relatedly, there is emerging evidence that it has become more important to citizens in recent decades that politicians appear more 'human' to them (Clarke et al., 2018;Garzia, 2011). As Clarke et al. (2018: 208) describe it: These authors found that citizens' anti-political sentiment has been rising steadily in the UK and that this has gone together with changing expectations of politicians: while citizens have always associated 'The Good Politician' with personality traits related to integrity and competence, there is a growing expectation that politicians should also be more 'human', 'normal' or 'in touch' with ordinary people: to be more authentic. Charisma is a term better reserved for when exceptional qualities of vision, veracity and trustworthiness are perceived in a leader by followers (Conger and Kanungo, 1987;Willner, 1984). Moreover, there is an important distinction between the trait of being able to persuade people on issues and being perceived as likable and like them. Indeed, it has been argued for a long time that with the advent of television and the rise of post-materialist values (Inglehart, 1997;Inglehart and Welzel, 2005), citizens have begun to place more importance on politicians appearing to be likable and similar to them, on the same level rather than above them (Garzia, 2011;Meyrowitz, 1985;Rahn et al., 1990).
The concept of 'authenticity' has a long and complex history within diverse academic fields: from the ancient Greek philosophers' preoccupation with the importance of knowing ourselves and expressing our true selves in life, through the existentialist writings of philosophers such as Jean-Paul Sartre and Michel Foucault about the difficulty of discovering and creating oneself (Daigle, 2017;Sutton, 2020), through studies in economics, gastronomy, music, nursing, tourism and various other fields of society (Catalano, 2000;Newman, 2019). In more recent times, studies in psychology and management have largely converged on three types of authenticity: historical consistency (having an authentic connection to a prior time/entity as claimed; e.g. an authentic Picasso), categorical conformity (conforming to the norms of a category as socially construed; e.g. authentic Italian food) and value consistency (the true external expression of an individual's or society's internal values; e.g. authentic artist) (Wood et al., 2008). It is this third category which is of most relevance to our discussion (and arguably to earlier philosophical treatments): individuals being true to their internal values in their expression, living in accordance with their inner selves rather than in line with external pressures and contexts. More specifically, we might think of authentic politicians as those who act genuinely like their true selves and like normal people, rather than in accordance with what is convenient and expected of them by the political context and by external pressures such as the media, campaign experts and political strategy. This trait has received little attention in empirical research in political science, however (Stiers et al., 2019).
To distinguish authenticity from integrity (associated with honesty, being true to principles, keeping promises) and competence (associated with skill, effectiveness, getting things done) it would seem appropriate to focus on the human dimension -on whether politicians are seen as in touch with ordinary people, accepting of themselves and others and able to understand everyday life. Do they engage with popular culture? Do they know how others live? Do they react with shared humour to situations? These qualities are not necessarily easy to find in elected representatives, in part, as Allen and Cutts (2018: 79) argue, because having the ambition to be a politician marks someone out: 'people who run for political office are strange, that is, they are unusual, abnormal, unlike most other people'. The pertinent differences extend beyond gender and class to personality features, where prospective politicians are more confident, more open to new challenges and more contented with disagreement and conflict. These personality features might comfortably go along with the performance of competence and even integrity, but they might not be the easiest starting point for showing authenticity. Yet authenticity may be a virtue that politicians increasingly need to demonstrate, in part as a response to rising distrust in politics. This issue is explored in the following section.

The Drivers of Political Trust: Process, Performance and Authenticity
Making trust judgements about politics and government is a demanding task. It is not surprising therefore that Whiteley et al. (2016) find considerable volatility in the public's assessment of the trustworthiness of British governments, notwithstanding an over-arching trend of decline in levels of trust since the 1990s. A new government or a change of government personnel can get an uplift in trust and then if a failure of process or performance becomes prominent (as, for example, in the case of the 2003 Iraq War) a loss of trust can be triggered and subsequently sustained. The public is trying to make trust judgements in a context of low information and uncertainty.
The challenge for citizens becomes clearer once it is recognised, following Hardin (2006), that trustworthiness is not about the moral standing of the actor or institution that is trusted or not. In what he calls 'an encapsulated interest account' the assumption is that trustworthiness is driven by a sense on the part of the citizen that the politician or government agency (the focus of their trust) has their interests at heart and will take them into account in their decisions. To establish trustworthiness, the citizen asks: Does the politician or government want to take our interests or welfare into account in their actions?
Trust judgements do not require perfect or even a great deal of information to be obtained by the citizen. Cues, hints, scraps of insight can be enough. Indeed, this caveat makes the possibility of political trust achievable rather than some utopian ideal. At the same time, however, errors of judgements may be common, either lacking trust in trustworthy agents, or expressing trust in agents who fail to act in the best interests of their citizens. One common shortcut used by citizens is to focus on individuals rather than institutions. Whiteley et al. (2016: 238) found that in their study 'evaluations of the Prime Minister of the day are likely to provide a powerful and easily used heuristic for determining whether the government can be trusted'. Citizens may be more comfortable about judging an individual politician rather than a government, since those are the criteria that they tend to use in much of their everyday lives.
The two 'classic' frames that citizens use to make trust judgements are focused on the processes used by governments to reach decisions and the performance or outcomes that they achieve. These two forms of judgement connect, respectively, to integrity and competence values associated with political figures. In making judgements it would be reasonable if these two forms of assessment for citizens were to a degree connected: if you get what you want in terms of outcomes it might increase your tendency to see the process as fair (Verba, 2006).
The process argument rests on the idea that if government treats people fairly when it makes decisions, then that is more likely to build political trust (Hibbing and Theiss-Morse, 2001). But demonstrating fairness is by no means an easy task. It could be established by a government agency through giving people the information they need, sharing facts and developments, being accessible and giving appropriate feedback. The issue is not whether that is done, but if it is perceived to be done. As Carman (2010: 747) comments, process 'matters for public evaluations of political institutions. Procedural justice scholars have demonstrated that in several contexts individuals' assessments of how 'fair' they believe a process to be will influence their willingness to accept outcomes'. But in most of these analyses the citizen has some form of direct engagement with the decision. An extension of this argument is that giving citizens a say in decision-making more directly through various forms of democratic innovation (participatory budgeting, citizens' juries, citizen polls) might also increase trust (Fishkin and Mansbridge, 2017;Geissel and Newton, 2012;Smith, 2009), but with the important caveat that these forms of engagement must be seen as fair (Carman, 2010). Again, to make a trust judgement in these cases involves citizens having some direct engagement with a democratic innovation.
Making a performance judgement is also challenging for citizens. As Whiteley et al. (2016: 237) argue: '(b)y implication a good performance should increase trust in government, since it implies that parties are successfully delivering on their promises'. But they go on and note how evaluations are often 'not so much on what is delivered, but rather on who is delivering it'. Citizens 'will opt for a "safe pair of hands" or a leader they think is competent and trustworthy and avoid a leader who they think is not up to the job'. Perception is, again, the key issue. And this in turn will be influenced by the partisan position of the assessor; if the party you voted for is in power, you are more likely to trust it to perform. Green and Jennings (2017) show how political parties come to gain or lose 'ownership' of issues, how they are judged on their performance in government across policy issues, and how they develop a reputation for competence (or incompetence) over a period in office. Their analysis tracks the major events causing people to reevaluate party reputations, and the costs of governing which cause electorates to punish parties in power. The challenge is that generally voters seem to be not very attentive to politics, and when it comes to retrospective judgement about the performance of government, they may lack the capacity to effectively allocate credit or blame to political leaders. Insofar as they do judge, those judgements are often quite myopic, based on the past few months before an election (Achen and Bartels, 2016).
Both process and performance evaluations demand from citizens considerable effort and engagement with politics and practices. In the absence of that effort, shortcuts of various types make trust judgements more likely to be less definitive, more partial, and epiphenomenal. But is there another option for the trust seeking citizen? Grimmelikhuijsen and Knies (2017), drawing upon an extensive body of studies of trust in public and private sector organisations, suggest three components of trust by citizens. Two are the now familiar competency and integrity aspects but a third is defined as perceived benevolence: the extent to which a citizen perceives the agent to care about their welfare and to be motivated to act in the public interest. There is a connection here to the earlier discussion of authenticity, defined as the capacity to be perceived to understand the concerns of citizens and share an understanding of their everyday lives. Again, as Hardin's (2006) definition suggests, the issue of trustworthiness is about judging whether someone has your interests at heart. The perceived authenticity of a political actor could be a reasonable proxy for such a judgement. When searching for relevant insights to make trust judgements, subject to limited time and effort, citizens may well turn to considerations of the authenticity of political actors, alongside their competence and integrity.

Research Design
The research design of our study rests on two surveys: one directed at citizens and one at political elites. The nationally representative survey of 1881 citizens was conducted by Sky Data (a member of the British Polling Council) between 20 and 30 October 2017. The survey of political elites was conducted by the authors, distributed to over a thousand politicians and journalists between 4 and 20 December 2017. Prior research finds that response rates for surveys of political elites are low and probably falling, in part because of time constraints on respondents, the status of such studies and because researchers have to typically make first contact instead of going through a survey organisation (Bailer, 2014;Walgrave and Joly, 2018). To collect responses, we compiled email addresses for 650 MPs at Westminster, 338 Lords, 73 MEPs, 129 Members of the Scottish Parliament, 58 Assembly Members in Wales, 83 Members of the Legislative Assembly in Northern Ireland and 330 journalists. We sent invitations and two reminders to each of these and managed to get 61 responses -47 from representatives and 14 from journalists. In line with prior surveys of elites, this is not a high response rate or a large enough sample to draw statistical inferences about either the population of representatives or journalists, but the responses do give an interesting insight into the priorities of the elites who responded to the survey, and a valuable comparison with citizen responses. Details of the coverage of the two surveys are provided in the online supplementary material.
In both surveys, respondents were presented with 21 items derived from previous research, indicating potential characteristics/traits of a good politician (Clarke et al., 2018) -such as 'is principled', 'works hard' and 'has had a proper job outside politics'. We asked: 'Where 1 is least important and 10 is most important, how important are each of the following to making a good politician?' and then listed the 21 different traits (the full list is reported in the online supplementary material). These scales are the basis of the dependent variables of the first part of our analysis but, as shown in the following section, the mean scores of these variables were highly clustered towards the top of the scale with little discriminating power between them. To disentangle this for statistical analyses, we create new variables of 'relative importance': the difference between the score each respondent gives to each item and the average score that respondent gave to all items, to indicate the relative priority placed by respondents on each trait. Based on these relative measures, we conduct principal-component factor analysis and reliability analysis to create three combined scales from the underlying patterns in the prioritisation scoring of these various items: scales for integrity, competence and authenticity.
In the citizen survey, we also asked various questions about trust and confidence in politics and government, mostly based on questions commonly used by researchers in the field of political support. These included a question asking, 'How confident are you in the ability of the UK government to address . . . ' (on a 5-point scale where 1 was 'not at all', and 5 was 'very') a range of (11) issues, with topics ranging from immigration to education and the environment. They also included several questions about attitudes to politics and politicians: 'How much do you personally trust the UK government?' (5-point scale from 'do not trust them at all' to 'trust them very much'); 'Do you think that British politicians are out merely for themselves, for their party, or to do their best for their country?'; 'Do you think that the standards of honesty and integrity of elected politicians in the UK have been improving in recent years, have been declining, or have they stayed about the same?' and 'In general, do you feel that the people in government are too often interested in looking after themselves, or do you feel that they can be trusted to do the right thing nearly all the time?' (more details on these questions and response options are presented in the online supplementary material).
In the analysis, we use these questions to fashion more parsimonious variables indicating political (dis)trust: one variable for the mean confidence respondents have in the UK government across the 11 policy issues, one binary variable for those who indicate distrust in the UK government, one binary variable for those who think that British politicians are out for themselves (as opposed to their party or country), one for those who think that politicians' standards have been declining in recent years and one for those who feel that the people in government are usually or sometimes out for themselves. We again use factor and reliability analyses to create a single scale of political distrust for these measures. In the case of both this scale and the personality trait scales, we weight each measure in the combined scale according to their loadings on the underlying factors, instead of simply summing them indiscriminately as a raw alpha-based scale would (see Boermans and Kattenberg, 2012). 1 To examine the relationships between these scales, background variables and vote in the 2017 general election and 2016 referendum, we use multivariate regression models. Finally, we present a descriptive exploration of the views of the elite respondents, comparing them with the views of citizens.

Results
The presentation of our results is divided into three parts. First, we analyse citizens' judgement of what makes for a good politician. Second, we compare this judgement with that of political elites. Finally, we explore the relationship between trust in politics and citizens' expectations regarding the good politician. Figure 1 presents the average importance score that citizen respondents gave to each listed trait of a good politician. These average scores are highly clustered towards the upper end of the scale (most traits score higher than 7 out of 10), as noted previously. Nonetheless, respondents ascribed the least importance to the 'traits' (or characteristics) of politicians dressing well and being presentable in terms of looks and voice, while the four highest rated traits are broad personality traits that all appear to relate to integrity: being honest, trustworthy and genuine and meaning what you say. Traits related to competence also seem rather highly valued (e.g. working hard, being level-headed, wise and strong) and the same goes for a few traits related to the 'authenticity' of politicians: especially those of understanding everyday life and being in touch with ordinary people.

Who Wants What From Politicians? The View of Citizens
To move from these descriptive summaries to statistical analysis, we start by eliminating the overall upward bias and clustering of the scoring for these traits, as well as the potentially biasing (in this context) effect of respondents using overall mental scales differently from one another. We do so by creating variables for each item that indicate how each respondent rated it compared with the average score they gave to all items (simply, subtracting this average from the score for each item). While this conceals the fact that some respondents generally rate traits lower, it focuses on the relative importance or prioritisation of politicians' traits for each respondent, which we are examining here.
We then run a principal-component factor analysis of these 21 transformed variables, with promax rotation (which allows factors to be correlated in order to approximate simple structure) to explore the dimensional structure of these different expectations (Table  1). We explored the data running factor analyses allowing for three, four, five and six factors in turn -all of these resulted in three major factors which substantively seem to reflect the three underlying trait dimensions suggested by our theory. These are all presented in the online supplementary material, but here we present and proceed on the basis of the five-factor structure which resulted in the most theoretically straightforward factor structure (the other versions, while indicating the same underlying dimensions, had overlaps between factor loadings which would cause undesirable multicollinearity between factors in our analysis). The results of this factor analysis are presented in Table 1, displaying all factor loadings above 0.3. These indicate that there are indeed three dominant dimensions in respondents' expectations of politicians' traits: at least six variables load onto each of those factors and their lowest eigenvalue is over 2, while only three variables  (Miller et al., 1986;Pancer et al., 1999), two of these factors seem to rather clearly reflect the underlying traits of integrity (being sincere, trustworthy, honest, genuine, principled and meaning what they say) and competence (being wise, level-headed, clever and working hard). The third dominant factor seems to clearly capture the 'authenticity' element noted in previous studies (Clarke et al., 2018;Garzia, 2011): understanding everyday life, being in touch with ordinary people and being comfortable with most people. Also note that in this factor, politicians' cleverness, looks, voice and clothes are perceived as less important; prioritising these traits has a negative loading on this dimension. The fourth and fifth factors seem to reflect traits which may be partially distinct but also overlap with the other factors both empirically and theoretically. Since they account for a much smaller part of the overall variation in response, we focus on the three dominant factors in our empirical analysis for parsimony.
To create combined scales reflecting respondents' different emphases on these different underlying traits, we combine the measures loading onto each respective factor, weighing them according to their heterogeneous contributions to that factor (multiplying their scores by their loadings on that factor) (see also Boermans and Kattenberg, 2012) and run reliability analysis of the scales created from these weighted variables. The results indicate a high reliability for the first factor, integrity (Cronbach's alpha = 0.78), and the second factor, authenticity (alpha = 0.65). Removing the measures of being 'clever' and 'comfortable with most people' improved the score of the authenticity scale to alpha = 0.74. This suggests that while the underlying trait measured by these different variables is substantively related to an emphasis on politicians being comfortable with most people, and a lack of emphasis on them being clever, these measures deviate more from measuring that underlying trait than the other measures do (likely because there are other underlying traits that are also related to that measure). Therefore, we do not include these two variables in the scalar measure but will remain aware that this scale also relates to those two traits. 2 The third factor, competence, had a significantly lower reliability score (alpha = 0.30), but we retain that scale since those measures all load onto the same factor, removing any of them would not improve the reliability of the scale and this factor is a theoretically important one (as confirmed by prior studies).
Looking at the varying importance placed on these different underlying traits by groups of citizens, Table 2 presents the regression models of each scale on gender, age, vote in the 2016 EU referendum and vote in the 2017 general election (with voting for remaining in the EU and for the Conservative Party as the reference categories). Looking at model 1, we see that the coefficient for gender is moderately significant (at the 95% confidence level): men appear to be slightly more likely to prioritise integrity traits, relative to other traits, than women. The same applies to 2017 Labour voters (when controlling for age, gender and referendum vote), who are more likely to prioritise integrity than those who voted for the Conservatives. Looking at model 2, older voters are slightly less likely to prioritise authenticity traits than younger voters but the strongest effects we find in these models are the partisan differences on this variable: Labour voters are substantially more likely to prioritise authenticity than Conservative voters, as are respondents who voted for other parties and those who did not vote. In model 3, we see that older voters are also less likely to prioritise competence and the same applies for Leave voters. Figure 2 plots the coefficients and their associated confidence intervals from these models, illustrating that the starkest dynamics appear to be the varying priority placed on authenticity by different groups of voters.

Elites in Touch but With Different Priorities
Turning to a descriptive exploration of the relative importance attributed to each scale in the small-N survey of political elites (61 respondents from a group of MPs, Lords, MEPs, members of devolved legislatures and journalists), Figure 3 compares the average scores given by citizens and elites to the personality traits used to create the scales in this analysis (in descending order within each group). It appears that citizens generally give higher scores to almost all traits, but it is difficult to parse out any systematic differences. The overall ordering seems to be similar but it is notable that the trait of understanding everyday life appears to be given much higher priority by citizens than elites, and this is indeed one of the traits that made up the authenticity scale in the citizen survey. Figure 4 disentangles this by graphing the average scores given to the items within each scale by citizens and elites -divided into political representatives (elected MPs, MEPs and members of devolved assemblies as well as unelected peers) and political journalists. Here, we see clear indications that authenticity is most valued by citizens, while integrity is most valued by representatives and competence most valued by journalists.  Calculating the same scales for the elite respondents (using the factor loadings from the citizen survey as weights), Figure 5 compares the average score on these scales by three groups: 1881 respondents from the citizen survey, 47 elected representatives from the elite survey and 14 journalists from the latter. Since these scales are partly based on factor loadings which differ between scales, these scores are comparable within scales but not between them (i.e. as we show above, citizens place more absolute importance on integrity than on authenticity). Comparing each scale between groups, we see further indication that citizens do indeed place the most relative importance on authenticity traits, more than the representatives and substantially more than the journalists from the elite survey, while the representatives prioritise integrity more and journalists place the highest relative importance on competence, by some margin.

Political Distrust and the Good Politician
In the first section of this analysis, we saw that when citizens rate the importance of different traits in making a good politician, there appear to be three major underlying traits that explain the relative priorities that they place on different traits: competence, integrity and authenticity. In the previous section, we then saw indications that citizens may place more priority on authenticity traits than political elites, while political representatives appear to place more priority on integrity traits and political journalists on competence. In this section, we turn to examining if and how these different traits are related to measures of political (dis)trust in the citizen survey. Starting with the different measures of trust available in the Sky Data survey, we have harmonised the variables so that all variables are made binary, where 1 indicates distrust and 0 indicates trust, except the measure of confidence in government; this is an average of 11 measures on a scale between 1 and 5, where higher values indicate more confidence. Table 3 presents a matrix of the Pearson's R correlation between these different measures (detailed in the online supplementary material), showing moderate-to-high correlations between all measures. A confirmatory principal-component factor analysis revealed that all measures indeed load significantly and substantially onto a single, underlying factor which presumably reflects respondents' overall political trust. The loadings of each measure on this common factor are presented in the final column of Table 3.
Applying the same scaling method as before, weighting the variables by their factor loadings, results in a scale which has a high alpha value of 0.72. This value would be lower if we dropped any of the variables, indicating that together they form a reliable scale which   Table 4 presents regression analyses of these measures and the composite distrust scale on gender, age, Leave vote in the 2016 referendum and party vote in the 2017 general election. Note that models 1 and 6 are linear regression models and the rest (models 2 to 4) are logistic, 3 and that the average of confidence measures is the only one relating to trust (or 'confidence') rather than lack of trust. From the models and coefficients presented in Table 4, it is clear that party vote has a strong relationship with all of these measures, while other predictors show few and weak relationships. This is hardly surprising: it is well known that identifying with a governing party has a positive effect on political support (e.g. Blais and Gélineau, 2007;Citrin, 1974;Marien, 2011). The Conservative Party was the party in office when the survey was conducted and voters of other parties all show significantly lower trust on almost all measures. In model 1, we see that the coefficient for voting Leave is moderately significant; when controlling for other factors (including 2017 vote), Leave voters actually have slightly more confidence in the government to address various issues. Conversely, in models 3 and 4 we see that this coefficient is positive: Leave voters are more likely to think that British politicians are 'out merely for themselves' and that people in government are 'too often interested in looking after themselves'. From models 2 and 4, it also appears that older voters are less likely to distrust the UK government and to think they are too interested in looking after themselves. Figure 6 plots the coefficients from these models and their associated confidence intervals, illustrating that gender and age are rarely significant (their confidence intervals cross the light-grey line indicating no effect) but that 2017 vote has strong effects in almost all instances. Turning back to the dimensions of the good politician, Table 5 estimates regression models for each dimension on the distrust scale, in addition to the demographic and political controls. Some interesting relationships are evident. Controlling for party preference, there is a strong positive relationship between political distrust and placing higher importance on politicians' authenticity, and a weak but significant negative relationship between distrust and the competence trait. Labour voters and non-voters are also significantly more likely to value authenticity more highly relative to other traits, controlling for their age and political distrust. These electoral dynamics of expectations of the good politician are illustrated with a coefficient plot in Figure 7, where the strong relationship between political distrust and prioritising the authenticity trait stands out. It is also interesting to note that including political distrust in these models substantially weakens the variation between voter groups found earlier in Table 2 and Figure 2; the only effect remaining is a mildly significant effect of voting for Labour. This indicates that the higher propensity of those who prioritise authenticity to vote for Labour, another party or not at all (as opposed to voting for the Conservatives) in 2017 was closely related to their higher levels of political distrust.

Discussion
We have analysed results from two surveys conducted in 2017: a nationally representative survey of the British public; and a small-N survey of political elites. These surveys focused on what characteristics or traits make for a good politician. Is it integrity, competence or authenticity that is most important? The citizen survey also focused on political distrust, allowing analysis of the relationship between distrust and expectations of the good politician. The analysis generated three main sets of findings. First, the traits regarded as important for the good politician varied by demographic and political characteristics of citizens. Younger people, Labour voters, and non-voters place more value on authenticity. Older people and men tend to place more value on integrity. Younger people and Remainers place more value on competence. Second, there are indications that citizens might place more value on authenticity than elites. Within our small-N sample of elites, political representatives tended to place more value on integrity and journalists more on competence. While our sample is too small to be considered representative, this finding provides indications which warrant further research on the topic. Third, citizens who place a high value on authenticity are more likely to express political distrust, while citizens who place a high value on competence are less likely to express political distrust.
In the rest of this final section, we discuss the second and third main sets of findings. They correspond to different parts of what we might call an 'authenticity gap'. There are indications of a gap between the extent to which authenticity is emphasised as an important trait by citizens and political elites. Citizens appear to place more value on authenticity than do elites. This could be linked to a desire of voters for representatives who share their social background (Heath, 2018). Furthermore, citizens who value authenticity highly are more likely to distrust politicians and the politics they represent. This suggests there is also a gap between the expectations that citizens have of authentic politicians and the inauthentic politicians they perceive to be offered by parties. What remains for discussion is the question of why these two parts of the authenticity gap might exist.
The gap between what citizens expect of politicians and what political elites expect of themselves makes sense in the context of existing research comparing understandings of good political behaviour. Allen and Birch (2015) found that understandings of political ethics vary between politicians and citizens. While the former tend to focus relatively narrowly on the financial conflicts of interest covered by institutional rules and codes, the focus of the latter is usually much broader and covers 'discursive integrity' -keeping promises and not giving evasive answers -in addition to financial integrity.
The research of Allen and Birch connects to a second point on the expectations gap between citizens and politicians. It is often said that generals always fight the last war, or politicians always fight the last political battle. For Allen and Birch, the last battle between voters and political elites in Britain was centred on the issue of politicians' integrity. There had been the period of 'Tory sleaze' in the early 1990s, followed by the MPs' expenses scandal of 2009. In this context, it is hardly surprising that elected representatives value integrity highly at present. The question is: are citizens still so concerned with integrity, or have they moved on to another point of focus, such as authenticity? Stoker (2017), reviewing the first hundred years of near-universal suffrage in Britain, argued that while class and then competence dominated political exchanges for most of the century, we have seen an emerging emphasis on a politics of resentment. This political style looks to blame others and identify enemies as a basis of garnering support. This shift is driven by economic change, social and geographical polarisation, and a cultural framing based on emphasising the gap between 'metropolitan liberal elites' and 'the left behind'. In this view, many citizens in the current period are looking for authentic politicians who understand their lives, concerns, and interests and will stand up for them. Politicians both within the mainstream and beyond have developed and exploited the politics of resentment, developing various versions of populist framings that place them on the side of the people against a failing elite. Indeed, as Joe Kennedy (2018) notes in no uncertain terms in his book Authentocrats, British politicians are by no means unaware of this desire for authentic politicians; whether their pursuit of it is genuine or not. For example, we have heard Labour's commitment to 'the many, not the few' while Boris Johnson and the Conservative Party framed the Brexit dilemma as the people against a 'Remainer' establishment. At the Conservative Party conference in 2016, then Prime Minister Theresa May contrasted the citizens of 'nowhere', who talk about being global citizens, against her preference for those citizens who share history, language and a common culture. A populist politics of resentment shifts the focus of citizens from competence and integrity to authenticity; speaking on behalf of ordinary people against the elite.
Turning to the second part of the authenticity gap -the gap between citizen expectations of authenticity and citizens' apparent judgements of inauthenticity, which appear to drive political distrust -this is consistent with our previous discussion of the drivers of political trust. Citizens make trust judgements using authenticity judgements as a short cut. Still, the question remains: why do citizens currently judge politicians to be inauthentic?
Here, the framework of demand, supply and intermediary factors is helpful (Hay, 2007;Norris, 2011). On the demand side of politics, citizens are thought to have become more educated, more secure in economic terms, and less deferent to authority figures (e.g. Dalton, 1984;Dalton and Wattenberg, 2000;Inglehart, 1977;Inglehart and Welzel, 2010;Norris, 1999). A part of this could be that citizens have come to expect politicians like themselves, as opposed to elites with superior knowledge and skills, and consequently, different lives, experiences and interests. Another demand-side factor of relevance may be what Sennett (1977) called 'the ideology of intimacy'. For Sennett, as societies like Britain's became more capitalist and secular, citizens traumatised by these changes withdrew from public life and increasingly focused on personality and self. They increasingly expected rewards of warmth, trust and open expression of feeling from all experiences. They progressively expected more authenticity from political engagement.
In terms of intermediary factors, it is thought that changes in the media sector (deregulation, new providers, increased competition) led to changes in media coverage of politics (shorter stories, more focus on partisan conflict and institutional failure, more opinion and instant reaction). These changes in turn led to changes in political communication (more stage-managed soundbites and photo opportunities), which in turn led to citizens feeling distanced from politicians (Clarke et al., 2018). Another intermediary factor of relevance here is what Hoggart (1957) called 'democratic egalitarianism'. He studied the uses of mass literacy by popular newspapers and magazines in 1950s Britain and found that such publications would flatter their customers by depicting 'the common man' or 'the little man' as a hero. Hoggart foresaw that citizens would gradually come to compare politicians to this heroic figure and find most politicians to be lacking in 'the common touch'.
Why should politicians be found lacking in this regard? Here, supply-side factors help to complete the picture. The concept of an emerging political class is a core idea as the tasks of government have got more demanding and complex (Borchert and Zeiss, 2003). The state has expanded and modernised, political occupations have been professionalised and financial rewards for politicians have been introduced and improved over time (pensions, allowances, salaries). Career opportunities for politicians have expanded, now including positions on parliamentary select committees, as elected mayors, in quangos, and in the devolved administrations. These developments are thought to have led to the decline of the amateur politician, who could be an aristocrat but could equally be a union shop steward, and the rise of the 'professional' or 'career' politician, who is commonly university-educated, middle class, and seemingly part of a homogeneous political class.
This concept of 'the political class' is an important one. Whether such a class exists from the perspective of data on occupational backgrounds or class consciousness among politicians, journalists, lobbyists and so on is open to debate. However, we follow Allen (2018) in thinking that what matters most is the clear existence of a narrative of the political class. This narrative is made up of three claims: (1) political elites all come from similar backgrounds (the claim of characteristic homogeneity), (2) political elites all think the same (attitudinal homogeneity) and (3) political elites all act the same (behavioural homogeneity). The key message for the present argument is that members of the political class are imagined to be similar to each other and different from ordinary people. This might explain why they are thought to be inauthentic and therefore untrustworthy.
Finally, it is the social background of politicians that links the two parts of the authenticity gap explored in this article. Many citizens perceive politicians to be out of touch with ordinary people; to lack understanding of their everyday lives and struggles. These citizens distrust what they perceive to be inauthentic politicians. At the same time, few politicians appear to recognise the importance of authenticity to British politics at the present moment. Sharing social backgrounds and political styles with their colleagues, they take the authenticity of politicians for granted. Why would they not? The problem is that many citizens value authenticity, perceive a lack of authenticity, and may come to perceive a failure on the part of politicians to even recognise the issue of authenticity.

Supplementary Information
Additional supplementary information may be found with the online version of this article. Contents

Notes
1. Robustness tests using raw alpha-scales yielded the same substantive results as reported here. 2. Including those two variables in the scales produces statistical relationships that are substantively the same and significant but weaker. 3. The confidence measure is an average of 11 scales for confidence in the UK government to deal with various issues, while the other trust measures have been recoded into a binary variable where 1 = distrust. The distrust scale is then composed of all of these measures, weighted by their loading unto their common underlying factor.