Intra-organizational trust in public organizations – the study of interpersonal trust in both vertical and horizontal relationships from a bidirectional perspective

ABSTRACT The purpose of this paper is to enhance our understanding of intra-organizational trust in public organizations by studying interpersonal trust in both vertical and horizontal relationships from a bidirectional perspective. Previous research has focused on trust at a single level of analysis, ignoring influences from other organizational levels, which has led to gaps in our understanding of trust. In addition, few studies take a bidirectional perspective where a trustor is simultaneously a trustee and vice versa. Through a case study, we contributed to filling this gap by studying the antecedents of trust – ability, benevolence and integrity.


Introduction
The issue of trust covers many research disciplines and levels of analysis (for literature reviews, see e.g. Fulmer and Gelfand 2012;Schoorman, Mayer, and Davis 2007;Lewicki, Tomlinson, and Gillespie 2006;Nyhan 2000;Rousseau et al. 1998). Reviewing the literature on trust in public organizations, we find mainly two branches of trust research (cf. Nyhan 2000;Cho 2008). The first is trust from an external environment perspective, i.e. interorganizational trust, where Bouckaert (2012) identifies three research orientations: societal trust in the public sector, public sector trust in society and trust within the public sector. The two latter orientations have received much less attention than the former (Oomsels and Bouckaert 2014). The second branch is intra-organizational trust, i.e. interpersonal trust within public organizations, where a small but growing number of researchers have begun to recognize the value of studying how trust is managed in public organizations (cf. Carnevale 1995;Cho and Park 2011;Cho and Lee 2011;Cho andPoister 2013, 2014). The two branches of trust research are connected. Some researchers (see e.g. Newell, Reeher, and Ronayne 2008; Cho and Ringquist 2011;Cho and Lee 2011) argue that understanding interpersonal trust within public organizations is a prerequisite for understanding external environmental perspectives of trust.
Thus, intra-organizational trust is not confined to the inside of public organizations. Rather, high levels of interpersonal trust are regarded as an important managerial resource (Cho and Lee 2011;Cho and Park 2011;Cho andPoister 2013, 2014) for building trust with other organizations and among citizens, as well as citizens' trust in public organizations to be managed in a trustworthy and effective way (Newell, Reeher, and Ronayne 2008;Cho and Lee 2011). For example, plenty of public organizations strive to engage in more entrepreneurial processes that require a higher degree of autonomy in order to deliver expected public services (Höglund, Holmgren Caicedo, and Mårtensson 2018;Höglund and Mårtensson 2019;Moore and Benington 2010). In this context trust becomes important as high levels of interpersonal trust do not need as extensive management and control practices (Cho andPoister 2013, 2014). High levels of trust make it easier for supervisors to delegate (cf. Behn 1995) and supervisors and subordinates are more willing to engage in unselfish behaviours (Albrecht and Travaglione 2003;Cho andPoister 2013, 2014;Jeffries and Reed 2000). In short, previous research has repeatedly addressed the importance of high levels of interpersonal trust to create a positive workplace environment, improve the performance of public organizations and build organizational trust in relation to other agencies and among citizens (see e.g. Newell, Reeher, and Ronayne 2008;Nyhan 1999;Cho and Lee 2011;Cho andPoister 2013, 2014).
This paper focuses on interpersonal trust in vertical or horizontal relationships within an organization (cf. Ellonen, Blomqvist, and Puumalainen 2008;Krot and Lewicka 2012). Previous research on interpersonal trust has mostly examined trust in vertical relationships (cf. Dirks and Ferrin 2001;Dirks and Skarlicki 2004); more precisely subordinates' trust in their superiors (Burke et al. 2007), while there are fewer studies of horizontal relationships involving trust among peers (Burke et al. 2007;Tan and Lim 2009). In an editorial, Fulmer and Dirks (2018) contend that the previous isolation of trust research at a single level of analysis, ignoring processes and influences from other organizational levels, creates gaps in our understanding of trust. Following on that, we argue that an integration of trust research across multiple levels in organizations is sorely needed (Fulmer and Gelfand 2012), including the study of trust in both vertical and horizontal relationships, which are seldom investigated in combination (Nyhan 2000;Cho and Park 2011). Studies examining trust in either vertical or horizontal relationships tend to ignore the fact that trust does not occur in a vacuum between subordinates and superiors or among peers (see e.g. Fulmer and Dirks 2018). Rather, trust in vertical relationships may influence horizontal relationships and vice versa (Cho and Park 2011).
Moreover, research discussing interpersonal trust often takes a unidirectional perspective, describing a trustor's perception of a trustee. This is questioned by Hasche, Linton, and Öberg (2017), who argue that we need to address trust in terms of a bidirectional perspective where a trustor is simultaneously a trustee and vice versa. Hence, so far previous studies have mostly regarded one party as the trustor and the other as the trustee, without reflecting on how the trustor simultaneously acts as the trustee of the other and how this will affect trust. Accordingly, the purpose of this paper is to enhance our understanding of intra-organizational trust in public organizations by investigating interpersonal trust in both vertical and horizontal relationships from a bidirectional perspective. Public organization, as used in this paper, refers to agencies, municipalities and organizations that are funded by tax money and act in a public sector context. We have not included research on state-owned enterprises (SOE). In relation to the purpose, we investigate the following research question: How do high and low levels of interpersonal trust influence vertical and horizontal relationships?
As in the overall field of trust research, public management studies of intraorganizational trust are mostly quantitative, based on extensive survey data (cf. Cho and Park 2011;Cho andPoister 2013, 2014), i.e. research characterized by static snapshots of trust at a single point in time, and are meant to test its relationship with hypothesized variables (Lewicki, Tomlinson, and Gillespie 2006). In other words, there is a call for qualitative in-depth studies (Smollan and Schiavone 2013), which we intend to answer by studying the Swedish Public Employment Services (SPES), one of the largest tax-funded central agencies in Sweden. In 2018, it had approximately 14,500 employees at 320 local employment offices across the country. The agency is accountable to the Swedish parliament and government who provide its mission, long-term objectives and tasks. SPES is governed by a board and managed by a director general.
For the last decade, SPES has received a very low ranking in surveys of citizens' trust in Swedish agencies (www.sifo.se). As a consequence, the government has explicitly stated that the agency needs to deliver better performance, quality and service to (re)gain trust (Appropriation Letter 2013SAPM 2015). The agency has been heavily criticized by citizens and the government for failing to fulfil its missions and goals by delivering inadequate services. Dissatisfaction has also been raised internally by the agency's managers and employees. To deal with the trust issues, SPES launched a new strategy in 2014 labelled 'the Renewal Journey'. The Swedish government also initiated a reform in 2016 to support government agencies in developing and implementing trust-based management control (SAPM 2016). The reform aims to develop a state of governance by balancing the need for control with trust in employees' knowledge and experience (Budget bill 2015/16). The reform is part of the government's work to create more efficient public agencies as well as greater benefits for citizens (Appropriation Letter 2016).
To sum up: First, this paper contributes to the literature on public management and interpersonal trust by analysing trust as high or low in vertical and horizontal relationships within an organization. Vertical relationships pertain to the relationships between subordinates and superiors, while horizontal relationships relate to relationships among peers. So far we have a gap in knowledge when it comes to our understanding of horizontal relationships. Second, the paper contributes by addressing the importance of a bidirectional perspective in studies of interpersonal trust, where subordinates, superiors and peers have double roles as both trustor and trustee. Third, it contributes by showing that trust levels in vertical and horizontal relationships are intertwined, where interpersonal trust in vertical relationships influences, positively or negatively, interpersonal trust in horizontal relationships. Fourth, it contributes with a qualitative case study including rich empirical descriptions of how a public organization manages interpersonal trust.
The article continues as follows: first, the theoretical background, framework, analytical tool and research design are presented. Then the analysis and conclusions are discussed.
Research on intra-organizational trust in a public management context Trust research most often involves two specific parties, the trustor and the trustee. Trustor is the trusting party, whereas trustee is the party that is to be trusted (Mayer, Davis, and Schoorman 1995). There are a multitude of perspectives on trust, resulting in a plethora of trust typologies (cf. Lewis and Weigert 1985;McAllister 1995;Smith and Lohrke 2008). Scholars thus present a variety of labels of the concept. However, most of the typologies consist of two or three dimensions: a rational economic aspect, a social aspect and in some cases a behavioural aspect of the relationship between a trustor and a trustee (Hasche, Linton, and Öberg 2017). Mayer, Davis, and Schoorman (1995) wrote one of the most frequently cited papers in the literature on trust. They define three antecedents of trust that seem to be more prominent than others in the definition of trust: ability, benevolence and integrity. These three antecedents of trust are also the most commonly recurring ones in the literature on intra-organizational trust in public organizations (cf. Cho and Park 2011;Cho and Lee 2011;Cho and Poister 2013), and they often appear as a set of considerations that explain a major portion of trust in another party in intra-organizational collaborations. Behn (1995), among others, identified trust as one of the most important questions in public management research. Similarly, Cho and Poister argue (Cho and Poister 2014) that trust is critical in public organizations, and that its importance is not limited to private sector organizations. In general, research shows that organizations with a high degree of trust are less dependent on management control practices and rules (Cho andPoister 2013, 2014). For example, when supervisors trust their subordinates, they are more willing to delegate authority, which may reduce micromanagement (cf. Behn 1995). Several studies (cf. Albrecht and Travaglione 2003;Cho andPoister 2013, 2014;Jeffries and Reed 2000) also show that people with a high level of trust in their colleagues or management are more cooperative and engage in unselfish behaviours, such as information sharing. Nyhan (2000) develops a framework of trust in public organizations by considering both antecedents and outcomes of trust in vertical relationships (subordinates' trust in superiors), where participation, feedback and empowerment positively affect trust, leading to higher levels of commitment and productivity. In a similar vein, Albrecht and Travaglione (2003) identify key antecedents and consequences of trust in vertical relationships (subordinates' trust in senior management), where the results show that effective communication, procedural justice, support and satisfaction with job security predict trust in senior management. Cho and Ringquist (2011) explore the trustworthiness of supervisors in relation to several organizational outcomes, such as employee satisfaction, cooperation and perceived work unit performance, by focusing on three antecedents that build trust: the competence, benevolence and integrity of supervisors. Similarly, Cho and Lee (2011) examine whether the perceived trustworthiness of supervisors in federal agencies has positive associations with employee satisfaction and cooperation within work units. Cho and Poister (2013) also investigate whether subordinates' perceptions of HRM practices, specifically autonomy, compensation, communication, performance appraisal and career development, are related to trust at three distinct levels of management: department leadership, leadership team and supervisor, in government organizations. In another article, Cho and Poister (2014) explore several antecedents to, and outcomes of, trust by considering the relationships among managerial practices, the three levels of management, teamwork and performance in a public organization. Lastly, Cho and Park (2011) are among the few that investigate both vertical and horizontal trust. They examine it in relation to the effect on satisfaction and commitment in different referents, looking at trust in both supervisors and peers. Prior research has mostly discussed trust in vertical relationships from the perspective of subordinates in relation to superiors. There has been little study of the reverse relationship (superiors' trust in subordinates). Investigating vertical and horizontal relationships in the same study is unusual. However, in this paper we study both vertical and horizontal relationships to examine trust between subordinates and superiors from both sides as well as trust among peers, i.e. in relationships between subordinates as well as between superiors. Thus, we're acknowledging a bidirectional perspective of trust, where the trustor is at the same time a trustee and vice versa (Hasche, Linton, and Öberg 2017).

Theoretical framework
Following previous research on intra-organizational trust in public management, we draw on Mayer, Davis and Schoorman (Mayer, Davis, and Schoorman 1995, 712), who define trust as: [. . .] the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform actions important to the trustor, irrespective of the ability to monitor or control the other party. Important to note, in relation to the definition of trust, is that the literature on interpersonal trust defines a party as a company, an organization, a group of superiors or subordinates, or individuals (see e.g. Mayer, Davis, and Schoorman 1995;Rotter 1967). In this paper we align with research that addresses interpersonal trust as between and within groups of superiors and subordinates. We use the three antecedents of trust proposed by Mayer, Davis, and Schoorman (1995): ability, benevolence and integrity, building on previous research on intra-organizational trust in public management (cf. Cho and Lee 2011;Cho and Park 2011;Cho and Poister 2013).
Ability-based trust Mayer, Davis, and Schoorman (1995) state that ability-based trust refers to the skills, competencies and characteristics of the trustee, i.e. that the trustee can fulfil his/her promises and obligations. The domain of the ability is specific, since the trustee may be highly competent in some areas, affording that person trust on tasks related to that area. In other areas, the trustee may have little aptitude, training or experience, making the person less trustworthy (Mayer, Davis, and Schoorman 1995). People working in an organization must trust each other's abilities to share knowledge in a competent way. Ability-based trust can be built relatively quickly, because it is not based on emotional interactions (Jeffries and Reed 2000). In this paper, ability is understood as the trustee's competence in his/her role (Cho and Lee 2011), where the trustor needs to perceive the ability of the trustee as positive for trust to exist.

Benevolence-based trust
Benevolence-based trust, on the other hand, is based on the perception that a trustee wants to do good for the trustor (Mayer, Davis, and Schoorman 1995). Krot and Lewicka (2012) suggest that benevolence-based trust is a willingness to consider the trustor's interests in a decision-making process, a willingness to act with consideration and sensitivity to the trustor's needs and interests as well as a willingness and desire to do favours for other members of the organization in working towards a common goal. Thus, in this paper benevolence-based trust is understood as the trustee's willingness to care and to act in the interest of the trustor rather than acting opportunistically (cf. Cho and Lee 2011). Thus, benevolence has to be perceived by the trustor for trust to ensue.

Integrity-based trust
Integrity-based trust is the perception that the trustee adheres to a set of principles that are considered acceptable behaviour by the trustor. Such issues as the consistency of the party's past actions, credible communications about the trustee from others, and the extent to which the party's actions are congruent with his/her words, all affect the degree to which the party is judged to have integrity (Mayer, Davis, and Schoorman 1995). Thus, this last trust antecedent includes reliability, fairness, justice and consistency (Fulmer and Gelfand 2012). Hence, integrity-based trust is understood as the trustee's actions being perceived as consistent and the trustee's words and actions being perceived as congruent by the trustor (Cho and Lee 2011).

Trust in vertical and horizontal relationships
Interpersonal trust can emerge in vertical and horizontal relationships between trustor and trustee. Vertical relationships deal with trust between subordinates and superiors, while horizontal relationships centre around trust between peers, both at the subordinate (employee/employee) and superior (manager/manager) level. Each of those can act as both trustor and as trustee.
In vertical relationshipsbetween superiors and subordinatesmanagers rely on employees to complete work tasks, where, for example, employee competence is an important element of managers' evaluation of their performance. Schoorman, Mayer, and Davis (2007) showed that the extent to which managers delegate responsibilities to employees depends on competence more than benevolence and integrity-based trust. However, Knoll and Gill (2011) found that employees who are treated fairly, with respect and dignity, perceive their managers as benevolent and, therefore, reliable and trustworthy (see also Krot and Lewicka 2012).
In horizontal relationships between peers (employee/employee), Tan and Lim (2009) showed that perceptions of benevolence and integrity-based trust were positively related to horizontal trust among peers, while perceptions of ability were not (see also McAllister 1995). Similarly, Krot and Lewicka (2012) found integrity-based trust to be the most important dimension of trust in relationships among peers. Horizontal relationships and trust in peers assume that co-workers will support their peers and will not take advantage of them by, e.g. withholding information. Co-worker trust also leads employees to act on the basis of faith in the words and actions of their peers (Ferres, Connell, and Travaglione 2004).

Analytical tool
The theoretical framework can be summed up in an analytical tool presented in Figure 1. Vertical relationships deal with trust in relationships between managers and employees, while horizontal relationships centre around trust between peers, both at the superior (manager/manager) and subordinate (employee/employee) levels. Accordingly, the interpersonal referents discussed in this papermanagers, employees and co-workerscan exist at different organizational levels. The co-worker dimension refers to members of an organization holding relatively equal power or authority level and with whom an employee interacts (Tan and Lim 2009). Co-workers exist at both a superior level, where various managers interact with each other, and at a subordinate level, where employees interact. In this paper, ability-based trust is interpreted as the trustee's competence in his/her role, benevolence-based trust is understood as the trustee's willingness to care and to act in the interest of the trustor rather than acting opportunistically, and integrity-based trust is understood as the trustee's actions being perceived as consistent and the trustee's words and actions being perceived as congruent (cf. Cho and Lee 2011).

Research design
Case study: the Swedish public employment service This paper is based on a qualitative case study, where SPES is treated as the case. There are different ways of conducting a case study (Ragin and Becker 1992); the most common approaches use the case study and its techniques as a method per se (cf. Glaser and Strauss 1967;Yin 1994;Eisenhardt 1989), or focus on the interpretative aspects, rather than the methods (cf. Stake 1995Stake , 1998. In this paper, we follow the latter approach. Drawing on Stake (1995), this study interprets the case of SPES to convey and conceptualize an understanding of intra-organizational trust in vertical and horizontal relationships. Through this interpretative case study approach, we try to address gaps in existing theory (cf. Siggelkow 2007 We selected SPES in line with the recommendations of Flyvbjerg (2006), who argues that when the aim is to achieve the greatest possible amount of information on a given problem or phenomenon, an atypical or extreme case often reveals more information. SPES can be viewed as an extreme case. As stated in the introduction, in 2014, the agency received the lowest ranking possible in a survey of the Swedish agency's reputation and trust in its work, performed by Research International Sweden (www.sifo.se). During the last decade, SPES struggled with very low levels of trust and was heavily criticized not only by politicians, government and citizens (see e.g. Appropriation Letter 2013; Budget bill 2015/162,014; SAPM 2015), but also internally by its own employees and managers for failing to fulfil its missions and goals (SPES 2015).
The new strategythe Renewal Journeythat was launched in 2014 is an attempt to introduce trust-based management control and an entrepreneurial spirit to offer and deliver relevant service to its customers. At the same time, it is an attempt to move away from the detailed, authoritarian micromanagement that resulted in a punitive work culture. By working towards changing the culture of the organization and introducing a new leadership philosophy based on the ideas of trust-based management, the agency hopes to (re)gain trust internally within the organization. For this new strategy, three core values were introducedprofessional, inspirational and trustworthy. The work to implement the new strategy is to take place between 2014 and 2021. So far we have studied the years 2014-2018.

Data collection
In line with an interpretative case study approach, we conducted interviews, observations and document studies (cf. Stake 1995Stake , 1998. We studied SPES retrospectively from 2014, and in real time from autumn 2015 to 2018. One hundred three interviews were held at different hierarchical levels between 2015 and 2018. Interviewees included the director general; top, middle and line managers at the superior level; and officers (i.e. case officers and specialists) at the subordinate level. Others, such as the board of directors and politicians from the government, were also interviewed. Table 1 gives a summary of the interviews. Some of the respondents, such as part of the top management team and the general director, were also interviewed annually, resulting in 89 unique respondents. The interviews lasted between 30 and 120 minutes, averaging 90 minutes, and were transcribed verbatim. They were conducted in such a way that we had the respondents narrate the meaning they ascribe to the renewal process in the context of prevailing values, practices, multiple perceptions and underlying structures. To gain an understanding of values in an organization, methods are required that allow interviewees to talk about their experiences and what these experiences mean to them. (Höglund, Holmgren Caicedo, and Mårtensson 2018). In 2017 we dedicated a significant amount of time to observing and participating in internal meetings, e.g. top management teams' weekly meetings (approximately 80 hours), and seminars and training programmes (approximately 140 hours). Detailed field notes were made at all the observations, and where possible we made recordings and verbatim transcriptions.
The purpose of the document study was mainly to increase our understanding of contextual factors, such as why the agency changed its strategy towards trust-based management, and to gain an understanding of the authorizing environment of the government, media and citizens. In line with this, we collected and studied such documents as strategic plans, annual reports, press articles and government reports.

Data analysis
We began the data analysis by coding our data. The coding process can be described as the precursor to analysis and is quite different from conducting the analysis itself, as the goal is not to find results but to find a way to manage a large amount of text (Höglund, Holmgren Caicedo, and Mårtensson 2018). Following this, we used the analytical tool presented in Figure 1. We used this tool as a first step to code the empirical material into the trust antecedents of ability, benevolence and integrity. Every instance of the transcribed text (interviews and observations) that could be sorted into one of the three antecedents of trust was copied into one document. Each author did this on their own, intending to be as inclusive as possible.
In the second step, we started to empirically analyse each antecedent of trustability, benevolence and integrity--to see what common statements recurred. When analysing and comparing the results from the individual interviews, we began to see what was a common view for each group of top, middle, and first line managers, as well as the case officers and specialists. Initially, each author conducted this step individually; this was followed by several group meetings to discuss our potential findings. At this point we also started to separate trust into vertical and horizontal relationships. In this way the analysis shifted towards the group of superiors and subordinates in vertical relationships and towards the peers in horizontal relationships, leaving the individual's own point of view in favour of the common view of the group. It is important to note in relation to this that the citations in the findings section where we discuss the case are selected to represent the overall view of, for example, the top managers and the specialists.
In the last step, based on the definitions of ability, benevolence and integrity-based trust, we interpreted the antecedents to trust as low or high in vertical and horizontal relationships. As an example, ability-based trust is related to the skills and competencies of the people in an organization. If the interviews and/or observations described significant problems with information sharing (a competence) between different levels of management as well as between managers and officers, we defined it as a low level of ability-based trust in vertical relationships. Also, important to note here is that we analysed the antecedents to trust as high or low by considering both the trustor's and trustee's perspectives. The results of this analysis are presented in the upcoming section.
Findingsthe case of the Swedish Public Employment Service

Ability-based trust in vertical and horizontal relationships
Ability-based trust is related to the skills and competencies of the people in an organization. Our findings show both low and high levels of ability-based trust in vertical and horizontal relationships at SPES. We exemplify this with a significant training initiative introduced at the agency to enhance managers' skills and competences as leaders.
In early 2015, the training programme was launched at various organizational levels aiming to develop behaviours related to the new strategy and to prevent them from falling back into old behaviours. A line manager stated the following: We use the waterfall approach. We work with all managers, all management teams and every level all the way down to change the way SPES manages its work. [. . .] Today we put a lot of effort into building up their trust [. . .]. I shouldn't be the one pointing and saying "do this and do that"; rather, they should make their own decisions and work based on them.
The training programme was conducted by external consultants, starting with top management, then middle management and finally the line managers. The approach built on top managers' trust in the ability of lower-level managers to act as change agents, and ability-based trust can be described as high in vertical relationships when discussing top management's trust in subordinate managers and the officers' trust in their superiors. However, it can also be described as low in vertical relationships when it comes to the skill of sharing information. One line manager stated: The waterfall method resulted in a lake at the middle management level. Perhaps a small trickle at our level, but when we pass it on to the case officers it's no more than a drop [of water].
What this quotation exemplifies is that when the information reaches the subordinate level in the hierarchy, there is not much information left. The poor informationsharing ability in vertical relationships has a direct effect on ability-based trust, which can be interpreted as low at both the superior and subordinate levels. A middle manager said: We as managers have a huge amount of information [. . .] We have to pick out parts to inform others because we do not have time to inform about everything. And we may not always have the right tools to provide the information, such as slideshows, movie clips or informational material.
Middle managers described situations where lots of information was received, but where the difficulties lie in deciding what information to share, with whom and when. Several of these managers also expressed concern about superior managers' lack of competence in handling information. One middle manager stated: The findings further show that line managers and officers are concerned with the accuracy of the information they receive, since the information from top management does not always reach lower hierarchical levels or tends to change as the information migrates downward. One of the line managers argued: There is a lot of information that disappears along the way. As with the vertical relationships already discussed, our results show that ability-based trust is both high and low in horizontal relationships as well. However, the case illustrates more examples of high ability-based trust in horizontal relationships than was the case when discussing ability-based trust in vertical relationships. We start by addressing the horizontal relationships between case officers and specialists as peers at the subordinate level. In general, specialists describe situations where low levels of ability-based trust are visible in relation to case officers and their competence. For example, the specialists express concern that the case officers lack competence about the resources available for jobseekers with special needs. A specialist argued: We view it as a lottery for our 'customers', the job seekers and employers [. . .]. Because your access to our resources depends on which case officer you meet.
The specialists in general show low ability-based trust when it comes to the competence of the case officers, but in specific collaborations the situation is the opposite. There are 'multi-competent teams', which consist of case officers and specialists working together with job seekers with a problem such as a hearing impairment. Within these teams, the case officers have been trained and informed about all resources available for these types of customers. Thus, specialists and case officers work closely together daily with the common responsibility of helping a specific customer group. One specialist in a multi-competent team said: [Instead of 60], we got one case officer who was incredibly competent and interested. S/he learned a little sign language and was good at reading sign language.
The findings further show that the lack of information-sharing ability in vertical relationships (which we described as contributing to low levels of ability-based trust in these relationships) resulted in high levels of ability-based trust in horizontal relationships. An example of this can be seen in the work of reorganizing the agency in line with the new strategy. Hence, there was little information regarding the reorganization to lower hierarchical levels, resulting in an organization where case officers worried, due to the lack of information. As a result, since middle and line managers were not able to give any answers, case officers approached their peers the specialists with their worries as one way of trying to cope with their uncertainty about the future. The specialists, such as psychologists and conversational therapists, who are used to listening to and talking with troubled people in their work roles, eased some of the case officers' worries. A specialist explained: It must be difficult as an employee to go and talk to their boss. [. . .] It will be colleagues [peers] that provide support.
In other words, the case officers perceive high ability-based trust in relation to the specialists in horizontal relationships.
In horizontal relationships between middle managers, ability-based trust can be viewed as high, especially in the management groups, where middle managers meet on a regular basis to share information and to support each other. A middle manager stated: [We] can support each other and I gain insight, and we have very low thresholds towards each other. It is quite a trusting relationship. I think I have a very good management team.

Benevolence-based trust in vertical and horizontal relationships
Benevolence-based trust focuses on the trustee's willingness to care and act in the interest of the trustor rather than acting opportunistically. The findings show low benevolence-based trust in both vertical and horizontal relationships.
In vertical relationships, the low levels of benevolence-based trust can be exemplified by how managers and officers have interpreted the new leadership philosophy of self-leadership. In many cases the new leadership philosophy has been interpreted in an opportunistic way, favouring individual managers' and officers' interest and selfdevelopment rather than the overall development of the agency. Part of the renewal process incorporates changing the managers' and officers' behaviours in relation to a new culture built on entrepreneurial processes of becoming proactive, creative and changeable. However, our findings show that self-leadership and the entrepreneurial processes came with the consequence that managers and officers believed they could do whatever they wanted. This can be exemplified by the following statement by an officer: This opportunistic behaviour that the quotation exemplifies recurred in all vertical relationships. Several similar statements were made where managers and officers described being too entrepreneurial in that they experimented in a way that went beyond the intentions of the new strategy and leadership philosophy. Moreover, those who said they could do whatever they want tended to forget that self-leadership ought to be acted on with the agency's assignment, mission and goals in mind (the trustor). For example, one top manager argued: We're talking about self-leadership and so on, which they [employees] interpret as freely choosing the work they do. [. . .] But when you act locally, you have to link up to the whole and realise that it isn't freely chosen work at allwe have a mission.
The overall strategy emphasizes the importance of entrepreneurial processes. However, as the quotation mentions, all work at SPES should be linked to the agency's strategy and mission. If self-leadership is interpreted in the wrong way and managers and employees behave opportunistically by working merely for their own unit or solely to realize their own goals, it negatively influences benevolence-based trust. One middle manager described the importance of embracing entrepreneurial processes, the scope for action given, and the necessity of testing different ways of doing things, while at the same time following the regulations: You have to follow the regulations and guidelines, but beyond that, your own initiative is crucial [. . .] You have to take your own responsibility [. . .] Take responsibility for your own actions and make sure to do things and not always wait for orders.
However, most of the line managers and officers do not actively refer to the fact that the new philosophy of self-leadership must be balanced with regulations and guidelines. Rather, they tend to embrace the interpretation of doing as one pleases or focus on their own development. As a result, benevolence-based trust can be interpreted as low.
Benevolence-based trust is also low in horizontal relationships at SPES, especially at the superior level. Many of the line managers invest hard work and effort to make individual units flourish, without thinking about what is best for all units. This was clearly shown, for example, when it came to newly recruited line managers. In 2015 and 2016, the agency recruited more than 200 line managers, many with experience from the private sector. These managers were described as acting opportunistically by putting their own units first, instead of relating the progress of their own unit to what is best for all other units, resulting in low levels of benevolence-based trust. One middle manager argued: It may not have been the brightest idea to bring in so many [newly recruited managers from the private sector] at once. As a government agency, we didn't really have the structure and preparedness for this. For example, what it means to be a government employee. That's led to quite a few challenging situations. It's a completely different kind of leadership. You can't say 'I do my own thing, as long as I achieve my results' and not care about good administration.
The quotation illustrates that externally recruited line managers seem to forget that good management is needed to perform well as a government agency, where good management means cooperating with co-managers at the same hierarchical level to get results, not just focusing on the results of their own unit. Such behaviour can be interpreted as opportunistic and negatively influence benevolence-based trust, where the work becomes self-centred and not benefiting the agency as a whole. In general, the middle managers tend to stress the importance of the line managers needing to be able to better cooperate in order to not only deliver the best quality to the agency's 'customers', but also internally, for the agency to succeed. One middle manager stated: One problem, though, is that several managers said they do not trust their peers at the same hierarchical level. One line manager said: We need internal collaboration, but we do not trust the other party.
Several managers at different hierarchical levels made similar comments. Many argued that instead of collaborating with other managers at the same hierarchical level, it is easier to do the work themselves since they believe they do the assignment better than their peers, which is not in line with the new strategy of trust-based management.

Integrity-based trust in vertical and horizontal relationships
Integrity-based trust focuses on the trustee's actions being perceived as consistent and their words and actions being perceived as congruent. Integrity-based trust is the only aspect of trust at SPES that is different on the vertical and the horizontal levels. In vertical relationships, the results show both high and low levels of integrity-based trust. In horizontal relationships, we could only find low integrity-based trust.
High integrity-based trust in vertical relationships can be exemplified by the director general and line managers, who are frequently described as acting consistently with the new strategy. Moreover, the findings illustrate that managers and officers who have embraced the entrepreneurial aspects of the new strategy tend to see an opportunity in doing something different and new, being spontaneous, innovative and testing the limits, behaviour that is consistent with the new strategy. To this end, several managers and employees have begun experimenting with various ways, as top management has encouraged variation and entrepreneurial behaviour. One line manager described the joy of having a manager that trusted him/her to test and do things in a different way: I like [my superior's] management because they relinquish control over our sections, letting us be innovative and try out new methods.
It is also through experimentation and testing that managers and officers found new ways of performing their work. One middle manager described it as follows: Now we're in this experimental world. [. . .] We try out new methods [. . .]. We will eventually have to consolidate, of course, but right now there has to be a certain amount of freedom to work out what is most optimal.
As the quotation suggests, the agency needs to consolidate and become more consistent over time. There is a consistency that is in line with the new strategy and culture that could be interpreted as a high level of integrity-based trust; however, it is also challenging integrity-based trust as the new culture supports doing things in different ways with different work practices that do not support consistency, which in the long run could contribute to lower integrity-based trust.
Regarding low integrity-based trust, our findings show that there is a lack of consistency in vertical relationships when it comes to what the general director states and what the middle management says. This can be exemplified by one of the line managers saying: I'm not hearing that the director general says the same things as our marketing manager [a middle manager] does.
Moreover, what top management says collectively and how they act as individual managers is not consistent. As a result, the expectations regarding performance and scope for action differ, which in turn is described as a lack of steering that creates an insecure organization. This can be interpreted as low levels of integrity-based trust vertically, between subordinates and their superiors. One top manager commented: I think that an ideal level of steering makes people feel secure, which allows them to blossom. They feel they have the faith and authorisation of management, which leads to them doing a good job.
There is also an inconsistency between what managers say and how they act regarding things like the new culture of becoming more entrepreneurial, which requires a greater scope for action. This inconsistency contributes to the reason why integrity-based trust is low in vertical relationships. The findings show that managers in general are positive to the idea of becoming more entrepreneurial and there were several statements about its importance at all hierarchical levels. However, when subordinates began to act more entrepreneurially, they tended to be stopped by the management level above. These inconsistencies between words and actions are especially visible at the top and middle management levels. One line manager argued: We line managers have been good at embracing the new culture, but the managers above us continue as usual.
Accordingly, the words and actions of middle management are perceived as inconsistent, resulting in low integrity-based trust in relationships between line managers and middle managers.
In horizontal relationships, our findings show that it is mostly at the superior level that the managers' actions are not perceived as congruent, as many of them have difficulty acting in the way that they agreed to during management meetings. Several times we observed that the top management meetings quickly reached consensus and agreement. But when we talked to the subordinate managers and the case officers, it emerged that several of the top managers did not act in the way they agreed to at the meeting. This can be interpreted as low integrity-based trust in horizontal relationships between top managers.
Line managers also sometimes had problems acting in line with what they decided at the management meetings. Thus, they showed low integrity-based trust in relation to each other. For example, a middle manager felt obliged to eliminate the management meetings with his/her line managers as they acted untrustworthy towards each other in the group. The middle manager explained: The line managers used to have their own management meetings, but I had to eliminate them. I could see that it was not working. [. . .] As the demand for results increased, I needed to step in and unite the team as we were heading in too many directions. They [the line managers] made several decisions, but they did not follow them. They have not been loyal to each other.
Thus, in horizontal relationships at the superior level, integrity-based trust is low, especially in relationships among top managers and in relationships among line managers.

Discussion of findings
Our results indicate it is important to study trust in different types of relationships (Krot and Lewicka 2012;Cho and Park 2011) and from a bidirectional perspective (Hasche, Linton, and Öberg 2017). As Krot and Lewicka (2012) argue, trust in vertical and horizontal relationships is different, where vertical trust is generally more complex than horizontal trust; vertical relationships come from a place of power and control that has a substantial impact on subordinates. Moreover, by breaking down trust into the three antecedents of ability, benevolence and integrity, we could engage in a more nuanced discussion of trust in relation to when it can be perceived as high or low. Mayer, Davis, and Schoorman (1995) argue that each antecedent provides a unique perspective from which to consider the trustee, while at the same time offering a conceptual framework for the empirical study of trust in another party. To sum it up, by studying the antecedents of trust as high or low, and its influence on vertical and horizontal relationships, our results show several interesting findings that we further elaborate on in this section of the paper. Table 2 summarizes the findings.
Ability-based trust is the only form of trust that is high in both vertical and horizontal relationships at the agency. This indicates that, as Jeffries and Reed (2000) argue, ability is one of the antecedents of trust that can easily be changed, as it is not based on emotional interactions. Moreover, the results also show that ability-based trust can be perceived as both high and low in vertical and horizontal relationships. Depending on what perspective we took on the trustor or trustee, we got different results. Thus, our results indicate that trust can be perceived as high or low depending on who takes the role of the trustor (trusting party) and the trustee (party to be trusted) in vertical relationships (Hasche, Linton, and Öberg 2017).
In the vertical relationship between top management and the line manager, trust could be perceived as high when it came to top management being the trustor and the line managers as the trustee. Hence, the choice of using the waterfall method shows that top management have great confidence in the competence of the line managers to train employees and act as change agents. In line with this, our findings support Schoorman, Mayer, and Davis (2007), who state that managers' willingness to delegate responsibilities to employees depends on competence more than benevolence and integrity-based trust.
If we reverse the role of the trustor and the trustee in the vertical relationship, the results show that ability-based trust can also be low. An example of low ability-based trust in vertical relationships is when it comes to the skill of sharing information. The case illustrates significant problems with information sharing, such as information overload, loss of information and poor accuracy of information received between line managers (trustor) and middle managers (trustee) as well as between officers (trustor) at the subordinate level and managers (trustee) at the superior level. As Jeffries and Reed (2000) state, managers and subordinates must trust each other's informationsharing ability to achieve high ability-based trust. Moreover, clear communication pathways are needed, with well-defined recipients of the information (cf. Albrecht and Travaglione 2003); otherwise uncertainty will arise, resulting in low ability-based trust (Cho and Poister 2014).
Interestingly, our findings show that low ability-based trust in vertical relationships due to poor information-sharing between managers and employees can positively influence ability-based trust in horizontal relationships. In line with Fulmer and Gelfand (2012), our findings show that trust in one referent acts as a substitute for trust in another referent to achieve desirable outcomes. Case officers turned to their peersthe specialiststo assuage their concerns about the reorganization, since middle and line managers were not able to give any answers. Accordingly, we argue that trust in vertical and horizontal relationships is an intertwined process, where trust in vertical relationships also influences relationships on the horizontal level (cf. Cho and Park 2011).
Moreover, in horizontal relationships it is also interesting to note that ability-based trust is perceived as low when specialists are trustors and case officers are trustees, but high when the roles are reversedcase officers as trustors and specialists as trustees. These findings support the importance of taking a bidirectional perspective on trust if we want to enhance our understanding of interpersonal trust and its effects on horizontal and vertical relationships (Hasche, Linton, and Öberg (2017). Benevolence-based trust is the only form of trust that is low in both vertical and horizontal relationships within the agency. For example, opportunistic behaviour indicates low benevolence-based trust (Mayer, Davis, and Schoorman 1995) and could explain the results. Hence the new leadership philosophy tends to generate interpretations that serve the managers themselves rather than the agency as a whole.
High benevolence-based trust in vertical and horizontal relationships is important as it involves supervisors treating subordinates with respect and fairness to encourage subordinates not to act in their own self-interest (Knoll and Gill 2011;Krot and Lewicka 2012). As a consequence, high benevolence-based trust in vertical relationships also becomes important for the horizontal relationships, as it will affect benevolence-based trust on a horizontal level. As Knoll and Gill (2011) state, employees who were trusted by their supervisors tended to also trust their peers, which is supported in our findings. Hence, our case shows how benevolence-based trust in vertical relationships engenders a lack of benevolence-based trust in horizontal relationships.
Apart from ability-based trust, integrity-based trust in vertical relationships, especially towards the director general and the line managers, is the only form of trust to show signs of high levels. It seems that these managers act consistently in relation to the new culture and leadership philosophy. In other words, despite its efforts so far, the agency still generally shows low levels of trust, especially in horizontal relationships. These results can probably be explained by the agency's history as an authoritarian culture promoting micromanagement, which previous research has shown has a direct negative impact on trust (cf. Behn 1995).
When it comes to horizontal relationships, integrity-based trust is low at SPES as employees demonstrate a lack of faith in the words and actions of their peers (cf. Ferres, Connell, and Travaglione 2004). This can be regarded as significant for the agency's continued efforts towards trust-based management control, as Krot and Lewicka (2012) found integrity to be the most important dimension of trust in relationships among peers.

Conclusions
This paper has resulted in four significant contributions. First, it contributes to the intra-organizational trust literature, by illustrating how the antecedents of trust function as high or low in vertical and in horizontal relationships. We do this using an analytical tool that focuses on the trust antecedents of ability, benevolence and integrity-based trust in vertical and horizontal relationships. Previous research has tended to isolate trust research at a single level of analysis, ignoring processes and influences from other organizational levels, which creates gaps in our understanding of trust (Fulmer and Dirks 2018). This resulted in previous research mostly examining trust in vertical relationships (cf. Dirks and Ferrin 2001;Dirks and Skarlicki 2004) and the subordinates' trust in their superiors (Burke et al. 2007).
Second, the paper contributes by addressing the importance of a bidirectional perspective in studies of interpersonal trust, where a trustor at the same time is a trustee, and vice versa. Previous studies have mostly regarded one party as the trustor and the other as the trustee, without reflecting on how the trustor simultaneously acts as the trustee of the other (Hasche, Linton, and Öberg 2017). Depending on who takes the role of the trustor and the trustee in the discussed relationships, the trustor can experience high confidence in the abilities of the trustee, but in reverse relationships with changed roles, the experience can change to low confidence in the counterpart's abilities.
Third, the paper contributes by showing that trust in vertical and horizontal relationships are intertwined, where interpersonal trust in vertical relationships can influence, positively or negatively, interpersonal trust in horizontal relationships and vice versa. Our results support that high trust in vertical relationships can positively influence trust in horizontal relationships. On the other hand, our results also support that lack of trust in vertical relationships can contribute to lack of trust in horizontal relationships.
Fourth, the paper contributes with a qualitative case study including rich empirical descriptions of how a public organization has managed interpersonal trust in vertical and horizontal relationships. Previous research in public management studies of intraorganizational trust are mostly quantitative, based on extensive survey data (cf. Cho and Park 2011;Cho andPoister 2013, 2014), resulting in a call for research using qualitative methods (Smollan and Schiavone 2013).
Limitations and suggestions for further research Mayer, Davis, and Schoorman (1995) argue that ability, benevolence and integrity are related, but at the same time separable. In line with this, our study shows that ability, benevolence and integrity vary independently of each other. The question is how low the levels of any of the antecedents can get before trust does not exist. It would also be interesting to further investigate in which specific situations each of the three antecedents are most sensitive or critical for trust to remain.
Furthermore, the results of our study showcase only that levels of trust in vertical relationships influence trust in horizontal relationships and not the other way around. In relation to this, we would like to encourage further studies of trust to embrace a view of vertical and horizontal relationships as intertwined and study how horizontal relationships could potentially influence vertical relationships.
Our study does not take the time perspective into account, which can be seen as a limitation. Hence, our results indicate the need for future studies that discuss the development of trust in terms of past, present and future. Based on these results, it can be argued that the evolution of trust over time in vertical and horizontal relationships depends on how the interacting parties manage the present, how they managed the past and how they will manage the future. This shows the importance of applying a processual approach to enhance our understanding of how trust is built and managed in public-sector organizations (cf. Höglund, Holmgren Caicedo, and Mårtensson 2018).

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by the Swedish Public Employment Service.

Notes on contributors
Nina Hasche (Ph.D.) is an assistant professor of business administration at Örebro University School of Business, Sweden. Hasche is a researcher at INTERORG Marketing Research Center and Center for Sustainable Business. Hasche focuses her research on relationship development both within and between organizations and she is a member of several research projects that take place within both public and private sector contexts. She is interested in change processes and collaborations between different types of actors in different contexts, where concepts such as trust, value and legitimacy are of high interest to her. She has published several papers on change, trust and value creation in different contexts. Nina Hasche is the corresponding author and can be contacted at: nina.hasche@oru.se.
Linda Höglund (Ph.D.) is an assistant professor at Mälardalen University School of Business, Society and Engineering (EST) in Sweden and a researcher at AES (the Academy of Management and Control in Central Government) at Stockholm University, Stockholm Business School (SBS) in Sweden. She is a project leader and a member of several research projects that takes place within a public sector context. Höglund focus her research on areas as management control from a strategic entrepreneurship perspective where she views strategic entrepreneurship as a process of organizing renewal in established organizations. She has written and published several papers on management control, strategy work, strategizing and strategic management in public sector, as well as on governance and board-work within a context of public sector. In relation to strategy and board work Höglund has also taken an interest in the study of trust and management of trust.
Maria Mårtensson (Ph.D.) is a Professor at School of Business and Economics, Linneaus University and Stockholm Centre for Organizational Research (Score), Stockholm University. Mårtensson is also one of the leaders of AES (the Academy of Management and Control in Central Government) at Stockholm Business School, Stockholm University which is a collaboration platform between state agencies and researchers. Mårtensson mainly research and publishing work within the field of management accounting and control, entrepreneurship in established organizations and strategic management. Mårtensson is a project leader of several major research projects that takes place within a public sector context. She has also written a large numbers of e.g. scientific articles, books and book chapters.