Leadership Perception in Remote Work Environments and Implications for Decent Work: An Inductive Qualitative Study in Indonesia

The organization of professional work has been fundamentally restructured over the past half decade. What the COVID-19 pandemic accelerated into emergency adoption has consolidated, across sectors and geographies, into a durable reconfiguration of where and how work is performed. In Indonesia, a nation of extraordinary organizational diversity spanning its high-growth technology sector, large public administration apparatus, rapidly expanding financial services industry, and deep higher education network, remote and hybrid work arrangements have become structural features of professional life rather than temporary accommodations to crisis conditions [1], [2]. The leadership challenge this creates is not merely logistical. It is epistemological: How do employees come to know their leaders when the ambient, co-present, relational texture of organizational life is systematically replaced by a reduced set of digital behavioral signals?

This question has received growing scholarly attention in Western organizational contexts. Research in Germany [3] and Europe more broadly [4], together with broad multidisciplinary syntheses of pandemic-era workplace research [5], has begun to identify the behavioral competencies and relational dynamics that distinguish effective remote leadership. Yet Southeast Asian organizational contexts—particularly those in Indonesia, with its distinctive combination of high power-distance culture, face-saving norms, and historically hierarchical institutional structures—remain substantially underrepresented in the empirical literature. This gap is not merely geographic. It is theoretical: The mechanisms through which power-distance dynamics shape how remote leadership behaviors are received, interpreted, and acted upon by employees constitute an explanatory blind spot that survey-based approaches, however large-scale, cannot adequately address.

The present study approaches this gap inductively. Rather than deploying an established leadership framework and testing its applicability in an Indonesian remote context, we begin with the accounts of employees themselves. We ask: What do employees actually perceive when they experience leadership at a distance? What micro-events change their sense of whether it is safe to speak, to raise a concern, or to acknowledge a mistake? What role does cultural positioning play not merely as a moderating variable, but as a constitutive element of the interpretive process through which leadership is experienced? Indonesia functions in this study not as a convenient research site but as a theoretically strategic boundary condition: a context where power-distance norms are sufficiently pronounced, and organizational hierarchies sufficiently institutionalized, that their interaction with remote work conditions becomes analytically visible in ways it might not be in lower power-distance settings.

The scholarly significance of focusing on employee perception rather than leader behavior requires explicit justification. Leadership research has predominantly examined what leaders do—behavioral styles, communication strategies, and transformational or transactional approaches—and assessed outcomes against leader-reported or observationally coded conduct [6], [7], [8]. Yet the pathway from leader behavior to employee outcome necessarily runs through employee perception and interpretation. Leaders who communicate one intent may signal something entirely different to employees whose cultural interpretive frameworks differ from their own. This gap between intended and received leadership signals is the core theoretical territory this study occupies, and it cannot be accessed through designs that capture only leader behavior or aggregate attitudinal outcomes.

This study is guided by a single, inductively oriented research question: “How do employees in Indonesian organizations perceive and interpret leadership behaviors and signals in remote work settings, and how do those interpretations shape their sense of safety and willingness to engage?”

The study makes contributions at three clearly differentiated levels. Empirically, it develops four themes that capture how employees experience remote leadership in a high power-distance cultural context. Theoretically, it derives three mechanisms—digital micro-behaviors as leadership signals, trust as asymmetric accumulation, and cultural permission structures—and integrates them into one process model in which trust and permission develop partly in parallel and jointly condition psychological safety and voice. Methodologically, it demonstrates the value of inductive qualitative inquiry for explaining interpretive and cultural processes that variable-oriented designs can identify only indirectly. Together, these contributions provide an employee-centric, process-oriented, and culturally situated account of remote leadership with implications for leadership development, organizational policy, and workplace equity.

The following review positions this study within four intersecting bodies of scholarship. Consistent with the study’s inductive orientation [9], [10], the literature functions here as a contextual framework for locating and elaborating analytically developed themes rather than as a deductive scaffold structuring the inquiry. Literature was engaged primarily after themes were constructed from the data, though the researchers’ existing theoretical knowledge inevitably shaped analytic sensitivity. This epistemological positioning is maintained throughout.

The review is organized around a specific diagnosis of the existing literature rather than a neutral survey of it. Three interrelated limitations recur across the bodies of work reviewed below. First, the literature is predominantly leader-centric: It theorizes the competencies, styles, and behaviors leaders should deploy while treating employees’ reception and interpretation of those behaviors as unproblematic. This is a theoretical problem because it presumes that communication is transparent that a signal sent is a signal received whereas digitally mediated settings systematically widen the gap between intended and received meaning. Second, the literature is predominantly variable-oriented: Cross-sectional survey designs establish that constructs covary but cannot specify the temporal, processual mechanisms through which trust, safety, and voice are constructed, maintained, or lost. Third, the literature is predominantly Western-centric, developed in low power-distance contexts whose interpretive assumptions about the availability of voice and the neutrality of hierarchy do not travel unexamined to Southeast Asian organizational settings. This study responds to that diagnosis by offering an employee-centric, process-oriented, and culturally situated account of remote leadership perception; each of the following subsections develops one strand of this diagnosis.

Physical separation between leaders and employees does not merely attenuate individual relational processes; it systematically alters the epistemological conditions under which leadership is perceived. In co-located settings, employees have continuous access to ambient relational information: the visible consistency between a supervisor’s public and private conduct, the spontaneous encounter that confirms availability, the non-verbal signal that a concern has registered. Remote work environments strip this informational layer away, replacing it with a deliberately constrained set of digitally transmitted behavioral signals that employees must interpret with greater intentionality and against reduced evidential baselines [11], [12].

The e-leadership and virtual leadership literature has produced two dominant explanatory models. A structural model foregrounds task interdependence, communication formalization, and team design as drivers of remote leadership effectiveness, emphasizing what organizational conditions leaders must establish rather than how employees come to know and respond to their supervisors [11], [13]. A behavioral-integrity perspective positions the perceived alignment between leaders’ words and deeds as a consequential basis for trust and credibility [14], while ethical leadership scholarship similarly emphasizes congruence between espoused values and enacted conduct [15]. These perspectives advance important insights, but they do not specify the interpretive process through which employees construct perceptions of leaders from the limited behavioral evidence available in digitally mediated settings. The limitation is not merely descriptive incompleteness. Existing frameworks often assume that leadership communication is received as intended, that interpretation is culturally neutral, and that the relevant repertoire consists primarily of intentional communicative acts, thereby overlooking the unintentional digital traces of latency, camera use, and channel selection from which employees also infer supervisory character.

More recent qualitative scholarship has begun to open this interpretive space. Rohwer et al. [3], drawing on 34 semi-structured interviews in Germany, identified trust-signaling and regular relational contact as the leadership behaviors most consequential for employee well-being in remote conditions, and argued that employees read these signals with a nuance and contextual sensitivity that quantitative instruments cannot capture. Kloepfer and Carbon [4], drawing primarily on expert and managerial perspectives, found that virtual leadership effectiveness requires consciously planned, empathic communicative strategies, and suggested that interpreting non-verbal and para-verbal cues demands greater attentional investment in remote than in co-located settings. Reviewing the e-leadership literature in the context of COVID-19 and beyond, Contreras et al. [16] argued that effective remote leadership requires distinct e-leadership competencies and a deliberate adaptation of communication practices rather than a simple transfer of co-located routines.

The post-pandemic consolidation of hybrid work has generated a second wave of empirical attention. Large-scale economic evidence indicates that hybrid arrangements have stabilized as a durable feature of professional labor markets rather than a transitional accommodation [17]. A randomized field experiment further found that hybrid working improved retention without damaging performance, reinforcing the view that hybrid arrangements can function as durable organizational systems rather than temporary concessions [18]. Kniffin et al. [5] identified remote leadership as a priority research area following their synthesis of pandemic-related organizational changes, noting that existing frameworks had been developed under assumptions of co-presence that no longer hold. Employee-focused evidence from the pandemic period further indicates that remote working effectiveness is shaped by work design and social conditions, including communication and monitoring practices, as experienced by employees themselves [19]. What remains substantially undertheorized is the employee side of this dynamic: The interpretive frameworks employees bring to digital supervisory behavior, the cultural schemas through which behavioral signals are decoded, and the temporal architecture of how trust and safety are constructed from an inherently impoverished behavioral data stream.

Psychological safety—the shared belief that interpersonal risk-taking within a team will not result in punishment or humiliation [20]—functions in this study not as a theoretical framework but as an outcome concept that participants consistently invoked in describing the consequences of supervisory behavior for their willingness to engage. Edmondson [20] foundational research identified leader coaching behaviors as a primary structural antecedent of team-level psychological safety. Subsequent meta-analytic work has confirmed leadership as among its strongest and most consistent predictors across organizational contexts [21]. Critically, these relationships appear to intensify under remote conditions: Mohase et al. [22] found that psychological safety mediated the relationship between inclusive leadership and employee voice among employees working fully or partly from home, providing direct evidence that the leadership-safety-voice pathway operates in digitally mediated work environments.

Employee voice—the willingness to offer ideas, raise concerns, and challenge the status quo through constructive communicative action—is both a downstream consequence of psychological safety and a crucial organizational learning mechanism [23], [24]. Its behavioral complement employee silence—the deliberate withholding of potentially useful information, concerns, or ideas—has received growing scholarly attention as a consequential organizational phenomenon in its own right [25]. In remote work contexts, silence takes on particular analytical importance because its presence becomes structurally less visible: Supervisors who perceive an absence of complaints may interpret that absence as satisfaction rather than suppression, misreading a product of constrained voice as evidence of positive engagement. Recent reviews confirm that psychological safety and employee voice remain central and actively developing constructs in contemporary organizational scholarship [26], [27]. Meta-analytic evidence establishes trust as consequential for the functioning of virtual teams [28], and planned, empathic supervisory communication has been linked to trust in virtual settings [4]; however, the pathway from such supervisory communication to employees’ willingness to speak up, and its cultural mediation, remains underspecified.

Indonesia constitutes a theoretically important context for studying remote leadership precisely because its cultural characteristics complicate the conditions under which psychological safety, voice, and trust are formed. Hofstede [29] cultural dimensions framework positions Indonesia among the world’s highest power-distance societies cultures in which the unequal distribution of power is accepted as legitimate, hierarchical relationships are deeply normatively embedded, and subordinates typically exercise strong deference toward superiors as a matter of socialized social order rather than strategic calculation. This cultural architecture creates conditions substantially different from the predominantly low power-distance Western contexts that dominate the remote leadership literature.

Himawan et al. [30] documented the sociocultural barriers to work-from-home arrangements in Asian high power-distance societies, arguing that remote work confronted many Indonesian organizations with conditions of interaction and self-management for which they had little institutional precedent. Sutarto et al. [1] found that enforced work-from-home conditions were associated with reduced job performance and satisfaction in Indonesia, with communication-related challenges and supervisory support among the factors associated with these outcomes. These findings suggest that leadership behavior carries heightened functional importance in Indonesian remote work contexts precisely because the institutional and relational infrastructures that normally supplement it are unavailable.

The concept of menjaga muka (face maintenance) in Indonesian organizational discourse carries specific implications for how critical feedback, public correction, and the expression of disagreement are experienced and managed. Botero and Van Dyne [31] demonstrated in a cross-cultural comparative study that power-distance orientation shapes employee voice behavior and conditions how leader–member relationships translate into voice, but variance-based designs of this kind cannot specify the mechanism through which such cultural conditioning operates. The present study is positioned to address precisely this explanatory gap, offering a process-level account of how cultural hierarchical norms shape the interpretive framework through which leadership signals are received and acted upon.

Trust in organizational contexts has been theorized across three developmental stages—calculus-based, knowledge-based, and identification-based [32]—reflecting different informational bases and relational depths. In remote work settings, the structural conditions for trust development are substantially altered. The ambient, informal behavioral data that sustains trust in co-located relationships—the visible follow-through on an informal commitment, the spontaneous expression of concern—is largely unavailable in digitally mediated environments, forcing employees to construct trust assessments from a more deliberate, constrained, and readily misread behavioral data stream [28], [33].

The meta-analysis by Breuer et al. [28] demonstrated that trust is positively associated with team effectiveness in virtual teams, with team virtuality and documentation moderating this relationship, a finding consistent with Sarker et al. [34] evidence that communication and trust jointly shaped performance in global virtual teams. What these accounts do not address is the temporal and asymmetric architecture of trust formation in remote settings: whether trust gains and losses operate symmetrically, how quickly positive signals accumulate relative to how rapidly negative events erode them, and how cultural interpretive frameworks shape the weighting employees apply to different behavioral evidence. These are the questions that the present study’s grounded approach is positioned to address.

The concept of digital communication competence has emerged as a distinct and consequential leadership capability in the e-leadership literature. Avolio et al. [13] proposed that e-leadership demands the development of qualitatively new competencies: the ability to read digital behavioral cues, calibrate presence across channels, and manage the cultural and relational implications of asynchronous communication. Liao [12] multilevel review extended this framework by emphasizing the distinct relational and communicative demands of virtual leadership, including the capacity to convey relational regard, trust, and accessibility through channels that strip out many of the cues through which these qualities are conventionally expressed.

Existing accounts have framed digital communication competence as a capacity that leaders intentionally develop and deploy. What they have not addressed is the substantial category of digital behaviors—response latency, status indicators, camera norms, channel selection—that function as leadership signals regardless of whether leaders intend them as such. This gap between intentional and unintentional digital signaling constitutes a key theoretical space that this study’s fourth theme addresses. The practical implication is consequential: Leaders who assume that digital communication behavior operates only when deliberately managed are systematically unaware of the continuous leadership signals their technical choices transmit.

This study adopts an interpretive qualitative design grounded in constructivist epistemology [10], under which meaning is constructed by participants and analytically reconstructed by researchers; accordingly, the study seeks contextually grounded accounts of how employees construct meaning from leadership behaviors encountered at a distance rather than universal mechanisms of remote leadership effectiveness.

The design is explicitly inductive: Analytic categories were allowed to emerge from sustained engagement with participant accounts, with theoretical frameworks engaged post-hoc as resources for locating and articulating empirically developed themes. The inquiry is theory-informed rather than theory-driven [9], and this distinction was maintained across all phases of the analysis.

Participants were recruited through purposive sampling designed to maximize diversity across sector, geographic region, gender, and organizational tenure, while ensuring each participant possessed direct, substantive experience of remote work with a reporting relationship to a supervisor [10]. Inclusion criteria required: (a) current employment in an Indonesian organization in a non-managerial or middle-management capacity (one participant, P48, holds a Dean position in higher education; this role was included on the basis that the Dean maintained a direct reporting relationship to a Rector throughout the remote work period, satisfying the subordinate-reporting criterion; it is acknowledged that a Dean’s positional resources differ from those of other participants, and P48’s account was accordingly read as the experience of a senior subordinate within a steep academic hierarchy rather than as equivalent to non-managerial experience, a caveat noted where P48’s data are drawn upon); (b) remote or hybrid work engagement for a minimum of six continuous months during or after the COVID-19 pandemic; (c) a direct reporting relationship to a supervisor during the remote work period; and (d) willingness to provide informed consent and participate in an audio-recorded interview.

The final sample comprised 48 participants across five organizational sectors—technology, financial services, higher education, government agencies, and professional services—and five major Indonesian regions: Java, Sumatra, Kalimantan, Sulawesi, and Bali. The deliberate cross-sector, cross-regional composition reflects the study’s commitment to analytic breadth: by distributing participants across substantially different organizational cultures and geographic contexts, the design enables identification of patterns robust enough to transcend individual organizational settings while remaining attentive to variation. Table 1 presents the full demographic summary.

Sample adequacy is assessed through the concept of information power [35] rather than the methodologically contested notion of saturation. Information power is determined by five factors: specificity of the research aim; specificity and relevance of the sample; degree of established theory in the area; quality of interview dialogue; and the analytic strategy employed. Against each criterion, this study’s design warrants a relatively small but analytically dense sample. The research question is specific and bounded, focusing on how a defined population experiences a defined phenomenon rather than pursuing organizational ethnography at scale. The sample is purposively constructed rather than convenience-drawn, with each participant selected to contribute a distinct sector, regional, and demographic perspective. Interview dialogues were extended (65–100 minutes) and in-depth, generating substantial contextually rich material per participant. The reflexive thematic analysis (RTA) strategy does not require formal saturation as an endpoint [36]. Across these dimensions, 48 purposively selected participants constitute a sample with strong information power for the aims of this study. Stated directly: 48 interviews were judged sufficient because the research question is narrow and bounded, the sample is purposively specific to that question, the interview dialogues were long and analytically dense, and RTA does not require saturation as a stopping criterion; by the final tranche of interviews, new transcripts were elaborating rather than altering the developing thematic structure. This assessment is consistent with recent methodological scholarship demonstrating that smaller, purposively designed samples frequently generate richer theoretical yield than larger convenience samples [35], [36]. The heterogeneity of the sample across five sectors and multiple regional locations across Indonesia is addressed through thick description and analytic generalization [37], rather than claims of statistical representativeness.

Data were collected through semi-structured, in-depth individual interviews conducted via Zoom or Google Meet, ranging from 65 to 100 minutes in duration. The interview guide was designed to elicit participant-generated meaning through experiential and reflective questions organized across six thematic entry points: (A) remote work background and experience; (B) leadership perception in digital settings; (C) trust and its development over time; (D) psychological safety and voice; (E) cultural and hierarchical dynamics; and (F) critical incidents and reflections.

The guide’s organization around recognizable entry points is acknowledged as a potential source of thematic pre-structuring. Three safeguards preserved inductive discovery: Questions were formulated experientially rather than conceptually, probes were non-directive, and coding proceeded across full transcripts rather than against the guide’s entry points, with several significant subthemes emerging from content that no question had anticipated. The residual limitation is discussed in Section 5.8.

Interviews were conducted in Bahasa Indonesia or English according to participant preference and transcribed verbatim. Indonesian-language interviews were translated by the first author and reviewed by the second author for cultural fidelity. Participant quotes used in the Findings section are presented in the language of the interview, with Indonesian terms retained and glossed where cultural specificity would be lost in translation. Member checking was conducted with eight participants after preliminary theme development [38], resulting in two instances of interpretive refinement: one participant clarified the role of collegial peer behavior in her permission-structure account, and one requested more precise differentiation between formal deference norms and personally held hierarchical values—a distinction that strengthened the analytic clarity of Theme 3.

Each main question was accompanied by non-directive probing prompts consistent with in-depth interviewing practice [39]. The full 20-question interview guide, organized across the six thematic sections described above, is available from the corresponding author upon request.

Data were analyzed using RTA [9], [36]. RTA treats themes as interpretive constructions produced through the researcher’s active, theoretically informed engagement with the data rather than as latent structures awaiting discovery; consistent with this epistemology, inter-coder reliability metrics were not used [36].

The six-phase process comprised: (1) familiarization through repeated reading and sustained reflexive memo writing; (2) inductive code generation at both semantic and latent levels, managed in NVivo 14; (3) candidate theme construction through systematic grouping and abstraction of codes; (4) theme review against coded extracts and the full dataset, with active negative case analysis; (5) theme definition, naming, and boundary clarification; and (6) narrative reporting with illustrative participant quotations attributed to de-identified codes (P01–P48). The coding process generated 87 initial codes, progressively abstracted into 24 focused codes, then organized into 9 candidate themes, and finally consolidated into 4 final themes with 12 subthemes. Table 2 presents representative codebook entries illustrating the coding progression.

Trustworthiness was established through Lincoln and Guba [38] four criteria. Credibility was pursued through prolonged and sustained engagement with the data, member checking with eight participants after preliminary theme development, peer debriefing with an external qualitative methodologist at two analytic transition points, and systematic negative case analysis at Phase 4 of the RTA process. Extended interview durations (65–100 minutes) enabled the development of contextually rich and internally coherent participant accounts that exceeded the depth typical of shorter qualitative protocols.

Transferability was supported through thick description of the study context, sampling rationale, participant characteristics, cultural setting, and organizational conditions. The appropriate standard for interpretive qualitative research is the analytic generalization—the application of theoretically elaborated concepts to comparable contexts—rather than statistical generalizability [37], and no claim of the latter is advanced. Dependability was addressed through a comprehensive audit trail documenting all analytic decisions across the six RTA phases, including rationale for code groupings, theme naming, boundary decisions, and the handling of discrepant cases. This trail is available for inspection upon request. Confirmability was established through reflexive journaling throughout the research process and through the positionality statement below.

Reflexivity is not a supplementary procedural step in RTA but a constitutive feature of the analytic process—the explicit, ongoing acknowledgment that the researcher is not a neutral conduit for data but an active co-constructor of the meanings that emerge from it [36]. The first author is a Bangladeshi researcher conducting organizational behavior research at an Indonesian university, occupying a position of disciplinary insider and cultural outsider simultaneously. This dual positioning carried both analytical resources and risks. As an outsider to Indonesian organizational culture, the first author approached participants’ accounts of power-distance dynamics, deference norms, and face-saving conventions as culturally specific phenomena requiring situated interpretation rather than transparent description. The first author’s direct professional experience in hierarchical, high power-distance Bangladeshi organizational contexts provided empathic access to the emotional textures participants described without generating assumptions of cultural equivalence—a distinction that proved analytically productive at several interpretive junctures.

The second author’s deep familiarity with Indonesian organizational life provided an insider counterweight, and analytic dialogue between the two authors was structured to exploit this complementarity. Points of interpretive divergence—particularly regarding the cultural specificity of the permission structure concept in Theme 3 and the appropriate analytic weight to assign cultural versus organizational-level explanations of hierarchical dynamics—were treated as generative analytical resources rather than methodological problems requiring resolution. Here is a specific source of potential interpretive bias: The first author’s prior theoretical engagement with authentic leadership and psychological safety frameworks may have disposed pattern recognition toward phenomena consistent with those conceptual territories. This risk was managed by deliberately delaying formal theoretical engagement until themes had been developed from the data, and by actively seeking accounts that resisted or complicated expected patterns. Two concrete instances illustrate how this reflexive discipline shaped the analysis in practice. First, an early candidate theme provisionally labeled in terms of authentic leadership was abandoned during theme development: On re-examination, the authors judged that the label imported the first author’s prior theoretical vocabulary onto the data rather than describing what participants were actually reporting, which was a narrower and more behavioral concern with consistency between statements and actions; the theme was accordingly reconstructed as behavioral consistency (Theme 1), a formulation that stays closer to participants’ own evaluative logic. Second, the two authors initially diverged on whether the deference patterns underlying Theme 3 should be interpreted primarily as expressions of national culture or as organizationally reproduced practices. This disagreement was resolved not by adjudication but by returning to the data, where the sharp contrast between government sector and technology sector accounts indicated that organizational and professional context conditioned the phenomenon at least as strongly as national cultural scripts—a resolution now reflected in the framing of Theme 3 and Section 5.3.1. The negative case analysis reported in Section 5.8 (particularly P14 and P06) similarly originated in a deliberate search for disconfirming accounts rather than emerging incidentally.

RTA of 48 verbatim interview transcripts yielded four overarching themes and twelve subthemes, presented in the following sections. The thematic structure represents an analytical construction built from sustained engagement with participant accounts rather than a discovered property of the data; the researchers’ interpretive agency is acknowledged at each stage. Table 3 summarizes the thematic structure with participant coverage. All four themes achieved broad participant representation, with participant coverage ranging from 36 to 48 participants. Theme 1 demonstrated the broadest representation, with evidence present across all 48 participant accounts. These coverage patterns establish the themes as robust and recurrent across the dataset rather than isolated or idiosyncratic accounts, while preserving space for the negative case analyses that qualify each theme’s scope.

The most consistently and emphatically foregrounded finding across all 48 interviews was that employees’ perception of their supervisors was organized not around formal leadership style categories but around the consistency or inconsistency between what supervisors stated about themselves—their openness, their commitment to the team, their accessibility—and what their behavior actually demonstrated over time. Participants described an active, ongoing inferential process: a sustained reading of supervisory conduct across interactions, channels, and time, through which behavioral signals accumulated into an increasingly confident assessment of whether risk-taking—speaking up, admitting uncertainty, raising concerns—was organizationally safe. This process was characteristically slow, extending over months rather than weeks, reflecting both the reduced density of available behavioral evidence in remote settings and the high stakes that erroneous trust assessments carry in high power-distance organizational cultures.

“My manager always said that she was open to feedback and that we were a team. But when I sent her a message pointing out a process problem, she did not reply for four days. Then in the next team meeting, she said nothing. After that, I stopped speaking up. The silence told me everything about whether my opinion actually mattered.” (P07, Financial Services, DKI Jakarta)

P07’s account is analytically representative in several respects. Trust was not broken by a confrontational event or explicit rebuke; it was eroded by the accumulation of a specific absence—the non-response—which, in the reduced informational environment of remote work, carried disproportionate interpretive weight. The absence of a response was itself read as data, as an implicit but unambiguous statement about the relational consequences of voice. This interpretive dynamic—the active meaning-making of supervisory non-action—recurred across participants in ways that co-located leadership research does not typically capture.

“What I noticed was that my supervisor admitted when he did not know something. During a project review on Zoom, he said openly, I am not certain about this direction, let us think about it together. That moment changed something for me. I felt like I could also say that I was uncertain without it reflecting badly on me.” (P03, Higher Education, East Java)

“At first I thought my manager was just being friendly in meetings. But after several months I started to see that she actually did what she said she would do. She followed up on things she promised. That consistency was what made me willing to bring problems to her rather than hiding them.” (P19, Technology, Batam, Sumatra)

The temporal dimension in P19’s account is theoretically significant: Trust was not formed from an initial positive impression but required months of behavioral accumulation before it became actionable as willingness to voice. This is consistent with P08 and P11’s accounts, which further specified that consistency across channels and in private as well as public settings was what distinguished trusted supervisors from those whose professed openness functioned as performance rather than practice.

“My previous manager was very enthusiastic in meetings. Always said, we are open to all ideas, come to me with anything. But in practice, nothing changed from those ideas. My current manager rarely says those things but she actually implements suggestions I raise. I trust the second one completely and the first one not at all.” (P08, Higher Education, Balikpapan, Kalimantan)

“I used to have a supervisor who was very warm in every one-on-one call. But then in the group channel she was completely different cold, directive, task-only. I never knew which version to trust. Eventually I trusted neither. My current supervisor is exactly the same person in every channel. That predictability is what I value most.” (P11, Technology, DKI Jakarta)

A subtheme of particular analytic importance concerned the interpretive significance participants attached to supervisory non-responses: delayed replies, absence of follow-through, and failure to acknowledge input. In remote work contexts, where positive behavioral signals already carry heightened interpretive weight given their reduced frequency relative to co-located settings, silences were consistently read as consequential negative data. The analytical implication runs directly contrary to the supervisory assumption that silence means nothing or merely reflects busyness.

“When I sent a proposal and heard nothing for two weeks, I did not assume he was busy. I assumed it was a signal. In the office you can walk past someone and read their mood. Online there is only what you send and what you get back. Silence is a reply.” (P04, Government Agency, DKI Jakarta)

The interpretive logic P04 articulates is organizationally consequential: Trust formation in remote settings is asymmetric. Negative signals register rapidly and erode the accumulated gains of consistent positive behavior. This mechanism is consistent with Baumeister et al. [40] general finding on negativity bias in social perception but has not been specifically theorized in the remote leadership literature. A single sustained silence can dismantle months of positive signal accumulation. The organizational implication is that supervisors who assume inaction is neutral are misunderstanding the communicative architecture of remote work.

Four participants offered accounts that complicated the theme’s formulation in analytically productive ways. P14 (Professional Services, Sumatra) described a supervisor who was behaviorally inconsistent by most criteria invoked by other participants—variable response times, different communicative personas across channels—yet was experienced as highly trustworthy precisely because the inconsistency felt authentic rather than strategic: ‘He is all over the place, but I always feel like he is being himself. The inconsistency is consistent, if that makes sense.’ This account suggests that employees may be engaged in a more nuanced interpretive project than ‘consistency detection’: They may be reading for authenticity for signs that what they are seeing is genuine rather than performed and under some conditions, the authenticity signal can be carried by behavioral variation rather than regularity. This complication does not undermine the dominant theme but specifies an important boundary condition: consistency is valued instrumentally, as the most reliable available signal of authenticity, not as an end in itself.

The second theme addresses a phenomenon that emerged with striking consistency across 37 of 48 participants: the heightened relational significance that supervisory disclosure of uncertainty, limitation, or error carried in remote work settings. Acts that in co-located settings might constitute ordinary professional candor took on amplified relational importance at a distance, where employees described feeling more isolated from organizational life and therefore more acutely responsive to signals that they were genuinely seen, included, and trusted. This interpretive intensification of disclosure functions not merely as a consequence of physical distance but as evidence of a structural asymmetry between leaders and employees in remote settings—an asymmetry that particular supervisory behaviors can either exploit or actively counteract.

“During the pandemic the company was struggling financially. My manager told us directly in a team meeting: I do not have all the answers right now, and I want to be honest about the uncertainty we are facing. He also said that he would protect us as much as he could. That honesty made me feel like I was not going to be blindsided. I felt safer because I was being told the truth.” (P02, Financial Services, West Java)

What is analytically significant in P02’s account is not merely the content of the supervisor’s disclosure but its function: in communicating genuine uncertainty, the supervisor simultaneously modeled the organizational acceptability of not-knowing—a modeling act that licensed employees’ own acknowledgment of uncertainty and limitation. This disclosure-as-modeling function recurred across multiple participants with remarkable consistency.

“There was a time when a project failed. My supervisor called me individually and said that she had made mistakes in the project planning. She did not blame the team. After that conversation I became more willing to admit my own mistakes, because I saw that it was acceptable.” (P15, Technology, DKI Jakarta)

“My supervisor once told us in a meeting that she had given the wrong advice to a client and that she was dealing with the consequences. I was surprised most managers would never say that, especially not remotely where everything feels more formal. But it made me feel that if I made a mistake I did not need to hide it.” (P13, Higher Education, DKI Jakarta)

“There is something about remote work that makes you feel more alone with your problems. When my supervisor acknowledged that he also struggled with the transition, that he also felt disconnected sometimes, it was the first time I felt that he understood what I was going through. That understanding made me want to perform, not just comply.” (P20, Professional Services, DKI Jakarta)

P20’s distinction between performance and compliance captures something that quantitative engagement metrics cannot easily access: the qualitative difference between motivated organizational participation and instrumental task completion. Supervisory disclosure, on this account, functions as a mechanism for converting compliance-based work relationships into engagement-based ones—a theoretically significant finding for organizational behavior scholarship on intrinsic motivation, identification, and psychological ownership.

Participants drew consistent and analytically important distinctions between supervisors who disclosed relevant information proactively before employees felt compelled to seek it and those who disclosed only when directly asked or when events made concealment untenable. Proactive disclosure was experienced as a relational investment that signaled the employee’s status as a person of worth within the supervisory relationship, not merely as a task-unit to be managed. Reactive disclosure, by contrast, was often experienced as insufficient and sometimes as evidence that the relationship was extractive rather than collaborative.

“When the restructuring news came, my supervisor messaged me before the official announcement. He just said, I want you to know what is happening before you hear it from someone else. That small act meant a great deal. I felt like I was seen as a person, not just a remote worker who receives information last.” (P10, Government Agency, West Java)

“She did not wait for me to ask whether my position was safe. She told me first. That made all the difference. In remote work, not knowing is worse than bad news. The information itself was less important than the fact that she thought to tell me.” (P05, Professional Services, Banten)

“There was a period when our department was undergoing evaluation. Other colleagues were asking around, anxious. My supervisor had already spoken to me. That preemptive honesty created a completely different relationship than I had with previous managers.” (P17, Financial Services, DKI Jakarta)

Several participants identified and articulated a structural feature of remote work that intensified the relational meaning of supervisory disclosure: The informational asymmetry between leaders and employees is systematically greater in remote settings than in co-located ones. In physical offices, employees passively absorb ambient organizational intelligence—the tone of hallway conversations, the visible configuration of closed-door meetings, the emotional weather of organizational life. Remote settings strip this ambient layer away entirely, making employees structurally dependent on what supervisors actively choose to transmit.

“In the office you know things without being told. You feel the mood. You see who is meeting with whom. At home you know nothing except what is sent to you directly. If your supervisor controls that information and shares little, you are completely in the dark. Some supervisors use that control deliberately. You can feel it.” (P16, Government Agency, West Kalimantan)

P16’s final observation that supervisors can and do use informational asymmetry as a deliberate relational instrument introduces a power dimension to the disclosure theme that has theoretical implications beyond individual-level trust. In high power-distance organizational cultures where informational advantage is recognized as a legitimate component of hierarchical authority, the decision to share proactively rather than maintain informational advantage represents a meaningful departure from culturally sanctioned norms. Supervisors who choose proactive disclosure in these contexts are making a culturally visible relational statement, which may account for part of the intensified relational significance participants attached to it.

The third theme is the most culturally specific and the one that diverges most sharply from existing remote leadership accounts developed in Western organizational contexts. Rather than simply transferring existing organizational hierarchies to a digital medium, participants consistently described remote work as producing an active cultural renegotiation of what hierarchical relationships permitted and prohibited—a process that operated differently across supervisors and organizations, unfolded over extended time horizons, and required specific behavioral evidence before its conditions could be changed. The default orientation toward supervisors that participants described was not individual wariness but culturally socialized deference: the normative expectation, internalized through years of Indonesian organizational and educational experience, that the appropriate relationship to authority is respectful compliance rather than candid exchange. An important interpretive caveat applies from the outset: Participants’ accounts reflect organizational and professional contexts at least as much as national cultural scripts, and the theme should not be read as a claim about a homogeneous Indonesian culture. The permission structures described below varied systematically by sector. Government participants (P04, P10, P16) described the most rigid and slowly shifting structures, whereas technology participants (P01, P11, P15) frequently described supervisors who had departed from hierarchical defaults well before the remote transition; higher education, financial services, and professional services accounts fell between these poles. Sector is therefore treated in this study not as a demographic descriptor but as an analytically consequential condition shaping the starting point and pace of permission accumulation, a point developed further in Sections 5.3.1 and 5.8.

“In Indonesian culture, you do not question your atasan [superior]. It is not how we are raised, it is not how the workplace works. So when my supervisor started actually asking for my opinion and then doing something with it, it felt strange at first. I had to learn that it was safe to speak. It was not automatic.” (P09, Professional Services, Makassar, Sulawesi)

P09’s account captures the process-level character of what this study terms cultural permission structures: the experience of needing explicit and repeatedly demonstrated permission before a culturally normative prohibition can be overridden. This process differs importantly from the trust-building process described in Theme 1, though the two are related: Employees must simultaneously assess whether they trust their supervisor (Theme 1) and whether the cultural prohibition on voice has been explicitly enough overridden that acting against it is safe. Both processes must be sufficiently advanced before voice becomes possible, and they accumulate on different evidence.

“There is always this question in the back of my mind: If I say this thing, will my supervisor think less of me? Will it affect my performance review? That fear does not disappear just because someone says they want open communication. You need to see it proved, many times, before you actually believe it.” (P16, Government Agency, West Kalimantan)

“My supervisor asked for our opinions in every meeting. At first I gave safe answers I agreed with what was already on the table. Only after I saw that other colleagues got positive responses for disagreeing did I try it myself. It took maybe six months before I said something that truly reflected what I thought.” (P18, Higher Education, Manado, Sulawesi)

“The hierarchy does not disappear in remote work. If anything it becomes more formal because everything is written down WhatsApp, email, formal meetings. In the office you can interpret a smile. Online you only have text and text carries status.” (P14, Professional Services, Palembang, Sumatra)

P14’s observation that text-based communication intensifies rather than attenuates hierarchical formality is theoretically significant and counter to the commonly held assumption that digital communication inherently democratizes organizational relationships. The inability to access non-verbal cues—the softening smile, the informal posture, the ambient warmth of physical proximity—meant that text communication between employees and supervisors often defaulted to its most formal, status-marked register. This textual formalization of hierarchy constitutes a mechanism through which remote work conditions can, paradoxically, strengthen rather than weaken power-distance dynamics at the level of everyday communicative practice.

Participants who had experienced critical feedback delivered in group video call settings described it as markedly more damaging to their willingness to participate, to their sense of relationship with the supervisor, and to their subsequent communicative behavior in team settings than equivalent feedback in physical environments.

“My previous manager corrected my work in front of the whole team on Zoom. In a physical office, it might have been manageable. But on video everyone can see your face. I was embarrassed in front of all my colleagues. After that, I never volunteered anything in team meetings again.” (P06, Technology, Medan, Sumatra)

“There is something about being on camera when you receive criticism. You cannot hide your reaction. Your face is right there in a little box for everyone to look at. A supervisor who corrects you privately treats you like a person. One who does it publicly, even gently, does not understand what that costs.” (P12, Financial Services, Surabaya, East Java)

These accounts illuminate a specific mechanism through which remote work amplifies face threats in high power-distance cultural contexts. The Brady-Bunch-style grid of a video call removes the recipient’s capacity to manage their visible emotional response through physical movement, spatial repositioning, or the strategic deployment of co-present informality. The involuntary public display of embarrassment—the loss of face captured simultaneously by every participant’s screen—is experienced as not merely equivalent to co-located face-loss but categorically more damaging because of its totality and its persistence in the memories of all those present. For participants in organizational cultures where face maintenance constitutes a primary social obligation, this structural feature of video communication represents a formidable barrier to the kind of honest corrective exchange that psychologically safe leadership requires.

Multiple participants described a distinctive temporal dynamic in the development of psychological safety under remote conditions: Safety was not formed through a decisive event or a single demonstration of supervisory benevolence, but developed slowly and hesitatingly through a sequence of low-stakes observational and participatory acts that incrementally extended the boundary of what employees perceived as safe to say.

“The first time I raised a concern in a remote team meeting I only said it because I saw a colleague do it first and nothing bad happened. I was like an observer for a long time before I became a participant. That observation period was much longer than it would have been in an office.” (P05, Professional Services, Banten)

“Safety is learned, not given. My supervisor can say he wants honest feedback but that means nothing until I have seen what happens when someone gives it. Remote work stretches that learning process because you have fewer data points fewer interactions to learn from.” (P10, Government Agency, West Java)

P10’s formulation safety is learned, not given encapsulates the process-level character of the permission-structure concept. The supervisor’s invitation to speak is not itself the permission; the permission is constructed from accumulated behavioral evidence of what actually happens when people accept that invitation. The reduced frequency of supervisory interactions in remote settings means that this evidential base accumulates more slowly than in co-located environments, extending the period during which employees operate under uncertain or prohibitive permission conditions. This has direct implications for how supervisors should design early-stage remote team interactions: creating deliberate opportunities for low-stakes risk-taking, and responding to them in ways that are unambiguously positive, can accelerate a process that might otherwise unfold over six months or more.

The fourth theme addresses a dimension of leadership perception that received minimal attention in the existing literature prior to this study’s development, yet emerged inductively with remarkable consistency: the way in which micro-behaviors of digital communication—camera usage patterns, response latency, channel selection, status indicators, message tone, and cross-channel consistency—were actively and consistently read by employees as continuous signals of leadership character, attentional presence, and authentic relational regard. The analytical significance of this theme lies not in identifying that digital communication matters to employees, which is established in the e-leadership literature, but in specifying the interpretive frameworks through which digital behaviors acquire relational meaning frameworks that operate largely outside supervisory awareness.

“A supervisor who turns off their camera in every meeting but requires us to be visible sends a message about hierarchy and visibility, whether they intend to or not. It feels like they are watching us without being watched themselves. That is not the dynamic of someone who is being transparent.” (P12, Financial Services, East Java)

“Response time matters enormously when you are working from home. If I message my supervisor about a problem and I wait three days for a reply, that delay tells me something about how I am valued. A supervisor who responds quickly, even just to say they have received my message, signals that they are present and that I matter.” (P17, Financial Services, DKI Jakarta)

“I have two supervisors one who uses formal email for everything and one who will message me on WhatsApp for something urgent. The WhatsApp supervisor feels more accessible, more human. I know that is just a channel choice but it matters. It tells me something about how much they think about me as a person versus a report to be managed.” (P01, Technology, DKI Jakarta)

The theoretical challenge these accounts pose is not merely that digital micro-behaviors matter, but that they function as relational signals regardless of the communicative intent behind them. P12 does not claim to know why the supervisor turns off the camera; the behavior is read and interpreted as a relational statement regardless of whether that reading corresponds to any supervisory intention. This gap between intended and received signals is a systematic and analytically consequential feature of remote leadership communication, with implications that the existing e-leadership literature, which predominantly frames digital competence as intentionally deployed does not adequately address.

“When my manager uses my name in a message, remembers something personal I mentioned, or responds outside working hours occasionally those things signal that I am not just a task to be managed. In remote work those signals are all you have. You read them carefully because they are the only material available.” (P11, Technology, DKI Jakarta)

Participants attended not merely to individual digital behaviors in isolation but to patterns of behavior across channels and contexts. Cross-channel consistency—being recognizably the same person in WhatsApp messages, Zoom calls, formal email, and group channels—was interpreted as a signal of authenticity with direct implications for psychological safety. The logic is straightforward: A supervisor who presents differently in different channels raises the interpretive question of which version is the genuine one, and the uncertainty this generates undermines the trust accumulation process described in Theme 1.

“There are supervisors who put emojis in messages and supervisors who write formally. Neither is better on its own. What matters is whether they are consistent. If someone is warm in a one-on-one but formal and distant in the team channel, I notice that. It suggests something is performed and something is not.” (P16, Government Agency, West Kalimantan)

A particularly generative subtheme concerned the extent to which technical and administrative choices—status settings on messaging platforms, notification practices, response time norms, out-of-office configurations—were interpreted as deliberate leadership signals regardless of whether they were intended as such.

“My manager always has his status set to ‘busy’ on Teams. I have never once seen him available. At first I thought it was technical. But after a while I realized it was a choice. It signals that he is not to be disturbed. That signal reaches me.” (P04, Government Agency, DKI Jakarta)

“She never sets an out-of-office reply, even on public holidays. I have sent messages on a Sunday and received a reply within twenty minutes. That behavior tells me something either that she has no boundaries, or that she genuinely cares enough to respond. I read it as care.” (P07, Financial Services, DKI Jakarta)

The interpretive work visible in P07’s account—the active meaning-making of an ambiguous behavior, resolved through attribution of positive intent—illustrates that employees are not passive recipients of supervisory digital behavior but active interpreters who construct supervisory character from whatever behavioral evidence is available to them. This has a direct implication for leadership theory: The study of remote leadership effectiveness cannot be conducted from the leader’s side of the relationship alone. The gap between what leaders intend and what employees receive is not a measurement error but an analytically central feature of the phenomenon.

The four themes constitute a coherent, empirically grounded account of remote leadership as experienced from the employee’s perspective in a high power-distance Indonesian organizational context. The central theoretical argument running through all four themes is this: Employees in remote settings are engaged in a continuous, active, and culturally mediated process of inferring supervisory character from a deliberately constrained behavioral data stream. What they are reading is not formal leadership style, stated commitment, or organizational policy, but the cumulative texture of behavioral signals transmitted intentionally or not across digital channels. The following discussion develops the theoretical implications of this account in relation to existing scholarship, advances three principal theoretical contributions, and addresses practical implications for leadership and organizational practice.

The data support a theoretically specified account of trust formation in remote settings as a cumulative, temporally extended, and asymmetric process. This extends Breuer et al. [28] meta-analytic demonstration that trust is consequential for virtual team effectiveness by specifying the temporal mechanism: Trust is not a function of any single communication act but of a pattern that accumulates across months of repeated behavioral observation. It extends Lewicki and Bunker [32] knowledge-based trust framework by showing that in remote settings, the knowledge base from which trust is constructed is a deliberately narrowed set of digital behavioral traces rather than the richly ambient relational data of co-located settings—a constraint that makes the knowledge-based trust stage slower to develop and more fragile once achieved.

The asymmetric character of the process—positive signals accumulate slowly, negative signals erode gains rapidly and disproportionately—is consistent with Baumeister et al. [40] general finding that bad is stronger than good in social perception, but has not been specifically theorized in the remote leadership literature. The present study locates this asymmetry in the specific context of supervisory communication behavior, specifying its most diagnostically powerful form: sustained silence in response to employee input. The practical implication is consequential and counter to intuitive supervisory assumptions. Supervisors cannot effectively offset a sustained communication failure by escalating positive signals; the asymmetric architecture of trust accumulation means that the damage a silence inflicts may substantially exceed the apparent magnitude of the failure, requiring disproportionately sustained repair effort. This is offered as an interpretive proposition for future longitudinal qualitative investigation rather than an established causal claim.

The remote leadership literature has converged on two broad explanatory perspectives: a structural perspective emphasizing task interdependence, formalization, and communication norms [11], [13], and a behavioral-integrity perspective emphasizing perceived alignment between leaders’ words and deeds as a basis for credibility and trust [14], [15]. These perspectives specify important features of what leaders do, but they do not fully explain how employees infer supervisory character from digitally mediated evidence, how those inferences accumulate or decay over time, or how cultural interpretive frameworks condition their translation into voice. The present findings extend prior work by specifying an asymmetric temporal architecture for trust formation, identifying digital micro-behaviors as a continuous and often unintentional signaling domain, and introducing cultural permission structures as a distinct process that interacts with trust in high power-distance contexts.

Theme 4 identifies a behavioral domain that the existing leadership literature has not systematically theorized but that emerges from participant accounts as central to how supervisory character is constructed and evaluated in remote settings. Camera usage, response latency, channel selection, status indicators, and cross-channel consistency are not peripheral communication choices; they function as a continuous stream of leadership signals that employees actively interpret as evidence about supervisory authenticity, attentional presence, and genuine relational regard.

The present data extend the e-leadership literature in a specific and theoretically important direction: Employees have developed informal but consistent interpretive frameworks for reading digital micro-behaviors as character signals—frameworks that operate continuously, often without supervisory awareness, and that are shaped by cultural expectations about what attentiveness, transparency, and respect look like in digitally mediated hierarchical relationships. This creates a systematic and consequential asymmetry between what leaders intend to signal and what employees actually receive. A supervisor who turns off their camera for entirely practical reasons may be unaware that this choice is being read as a statement about hierarchy and visibility. A supervisor who maintains a ‘busy’ status to manage interruptions may not register that this setting is being received as relational unavailability. The implication for leadership theory is direct: Digital communication behavior constitutes a leadership channel that requires deliberate cultivation. Leaders who leave it unmanaged are not communicating nothing; they are communicating without awareness or intention.

The concept of digital micro-behaviors as leadership signals overlaps with, but is not reducible to, several established constructs in the communication and leadership literatures. Table 4 situates the present study’s contribution against communication quality, signaling theory, media richness theory, and leader-member exchange (LMX).

The concept of cultural permission structures represents the study’s most process-specific theoretical contribution and the one that diverges most sharply from how power distance has been theorized in existing quantitative leadership research. The data suggest that in high power-distance Indonesian organizational contexts, the relationship between leader behavior and employee voice operates through a qualitatively different process than existing theory describes. Employees do not begin from a position of provisional openness, with power-distance norms moderating how strongly supervisory behavior shifts them toward voice. They begin from a default of hierarchical prohibition—a deeply socialized expectation that deference is the appropriate relational posture—and must accumulate explicit, repeated, and credibly demonstrated behavioral permission before that default can be overridden. This mechanism is not captured by existing power-distance moderation accounts [31], which treat power distance as modifying the strength of a relationship between variables rather than as inserting a qualitatively distinct accumulation process between leader behavior and employee response.

The concept is offered as a grounded description of how 46 of 48 participants described their experience, not as a generalized causal model applicable across all high power-distance contexts. Whether the permission-structure mechanism operates similarly in Malaysia, South Korea, China, or the Philippines is an empirical question for comparative qualitative research.

The cultural permission structure described in Section 5.3 is presented as a pattern observed among 46 of 48 participants, not as a property of ‘Indonesian culture’ conceived as a homogeneous national entity. Indonesia comprises substantial regional, ethnic, linguistic, and organizational diversity, and the sample spanning Java, Sumatra, Kalimantan, Sulawesi, and Bali across technology, financial services, higher education, government, and professional services was constructed precisely to surface variation rather than to assume uniformity. That variation is analytically visible in the data: Government sector participants (P04, P10, P16) described the most rigid and slowly shifting permission structures, consistent with the more hierarchical institutional norms associated with Indonesian public administration. Technology sector participants (P01, P11, P15, P19) more frequently described supervisors who had already substantially departed from default deference patterns prior to the remote transition, suggesting that organizational culture itself—shaped by sector, founding history, and workforce demographics—conditions the starting point of permission accumulation at least as significantly as any national-level cultural script.

Hofstede [29] power-distance framework remains a useful orienting reference for why a permission-accumulation process exists at all in this dataset, but it operates here as a backdrop condition rather than a proximate explanatory variable. The permission structures described by participants are better understood as organizationally and professionally situated practices reproduced, attenuated, or contested within particular workplaces than as direct expressions of a national cultural trait. Claims in this paper about cultural permission structures should accordingly be read as claims about the organizational and professional contexts represented in this sample: urban, knowledge-sector, and professional-services-adjacent workplaces across five major island regions.

Psychological safety [20] describes an experienced outcome state—the team-level belief that interpersonal risk-taking will not be penalized. Cultural permission structures describe a culturally conditioned mechanism through which employees approach that state. The present data do not support a strict sequence in which permission must be completed before trust can develop. Instead, trust and permission accumulate partly in parallel from overlapping but non-identical behavioral evidence: Trust concerns the supervisor’s perceived reliability and benevolence, whereas permission concerns whether upward voice is culturally and organizationally legitimate. Their interaction determines whether psychological safety becomes actionable as voice.

Voice climate [23] describes shared contextual perceptions of whether speaking up is encouraged, tolerated, or suppressed. Cultural permission structures describe the individual-level and temporally unfolding process by which an employee navigates that climate when the cultural default is hierarchical deference. An organization may possess a nominally open voice climate while employees still require repeated behavioral evidence before they can act within it. LMX theory [41] describes the quality of differentiated dyadic relationships that develop through reciprocal exchange; cultural permission structures instead concern the culturally shaped legitimacy of initiating substantive upward exchange. Trust development [32] and cultural permission structures are therefore analytically separable but mutually reinforcing. An employee may trust a supervisor yet remain uncertain whether candid voice is culturally permissible, or may receive nominal permission to speak without trusting that the supervisor will respond reliably.

This study advances one overarching theoretical contribution—an integrated process model of remote leadership interpretation composed of three analytically distinct but interacting mechanisms. Digital micro-behaviors as leadership signals constitute the informational input. Trust as asymmetric accumulation constitutes a temporal processing mechanism. Cultural permission structures constitute a culturally conditioned gatekeeping mechanism. After employees interpret leadership signals, trust and cultural permission develop partly in parallel rather than in a strict sequence; together, they determine whether psychological safety and voice become available.

This formulation clarifies why the same episode may inform more than one mechanism. Camera asymmetry is first encountered as a digital signal, then interpreted as evidence relevant to trust and to the legitimacy of upward voice. Likewise, supervisory silence can erode perceived reliability while also communicating that voice is unwelcome. Table 5 differentiates the mechanisms by antecedents, operation, and proximate outcome while preserving their interaction within one process model.

Figure 1 summarizes the integrated process model. Remote leadership behaviors generate digital signals that employees interpret; the resulting evidence feeds two partly parallel and interacting mechanisms trust accumulation and cultural permission accumulation which jointly shape psychological safety, employee voice, and engagement.

Trust and cultural permission are analytically distinct but mutually reinforcing. Trust concerns perceived supervisory reliability and benevolence, whereas permission concerns whether upward voice is culturally and organizationally legitimate. Both develop through repeated behavioral evidence, and neither is assumed to strictly precede the other. The model is presented as an interpretive proposition grounded in 46–48 participants’ accounts rather than as a tested causal sequence.

The present findings are substantially consistent with Rohwer et al. [3] German interview-based study in identifying trust-signaling behaviors and regular relational contact as central to remote leadership effectiveness, and with Kloepfer and Carbon [4] finding, derived primarily from expert and managerial perspectives, that virtual leadership effectiveness requires consciously planned, empathic communicative engagement. However, the present findings diverge from these accounts in cultural specificity: Rohwer et al.’s participants described trust primarily in terms of individual supervisor characteristics without invoking cultural prohibition structures. This divergence is theoretically meaningful rather than incidental: The permission-accumulation process may be substantially more salient in high power-distance contexts than in the low power-distance German settings of Rohwer et al.’s study, suggesting that the cultural context constitutes not merely a moderator but a constitutive element of what remote leadership effectiveness requires and how it is experienced.

The findings are consistent with Mohase et al. [22] finding that psychological safety mediates the relationship between inclusive leadership and employee voice in remote settings, while providing qualitative texture to the mechanisms underlying that mediation. What Mohase et al.’s survey design registers as an inclusive leadership effect may be better conceptualized, in high power-distance contexts, as a permission-accumulation process in which behavioral inclusion gradually overrides a culturally instituted hierarchical prohibition rather than simply strengthening an already available voice disposition. The study is partially inconsistent with accounts in the virtual-team literature that foreground transformational leadership as a particularly promising style for virtual contexts [12]. Participants did not describe their most trusted supervisors in transformational terms; behavioral consistency, not inspirational vision, was the primary trust-generating mechanism they identified.

The theoretical claims advanced here carry different boundary conditions, and these should be stated explicitly. Two mechanisms are proposed as candidates for cross-context transferability: the asymmetric temporal architecture of trust accumulation, which builds on a general negativity-bias regularity in social perception, and the operation of digital micro-behaviors as an always-active signaling domain, which follows from the structural signal impoverishment of remote work rather than from any culturally specific interpretive scheme. Both are nonetheless expected to vary in intensity with platform environments and organizational communication norms. By contrast, the cultural permission structure mechanism is proposed as strongly context-dependent: It presupposes a socialized deference default that is characteristic of high power-distance organizational settings, and the government-versus-technology sector contrast within this sample indicates that even within Indonesia its strength varies with organizational and professional context. Researchers applying these concepts in low power-distance settings should expect the first two mechanisms to be recognizable and the third to be attenuated, transformed, or absent.

The decent work framing in the title is grounded in the findings as follows. Decent work, as articulated in the International Labour Organization’s Decent Work Agenda [42] and in Sustainable Development Goal (SDG) 8 of the United Nations 2030 Agenda [43], encompasses not only employment and income security but also dignity at work, fairness of treatment, and meaningful participation in the decisions that affect working life. The mechanisms identified in this study bear directly on three of these dimensions in remote settings. Participation: The cultural permission structure functions as a systematic constraint on employees’ access to voice, and voice is the primary channel through which remote employees participate in organizational decision-making; where permission is never credibly granted, participation is formally available but substantively foreclosed. Dignity and fairness: The informational asymmetry documented in Theme 2 determines whether remote employees learn of decisions affecting them before or after those decisions are settled, a distinction participants experienced as a marker of respect versus disregard; similarly, the delivery of critical feedback through public versus private channels (Section 4.3.1) directly implicates dignity under face-saving norms. Quality of working life: The asymmetric trust dynamics of Theme 1 mean that sustained supervisory silence degrades not only trust but the experienced security of the employment relationship itself. Remote leadership practice, on this account, is not adjacent to the decent work agenda; the behavioral and system-design conditions specified in this study are among the proximate determinants of whether remote work is experienced as decent work. This connection is developed here rather than treated as incidental framing.

For organizational leaders and managers, the findings suggest a set of leadership behaviors that employees in remote settings perceived as consequential, behaviors that differ substantially from those emphasized in co-located contexts and that merit deliberate cultivation. These include: responding to employee communications with appropriate speed even when a substantive response is not yet possible, thereby counteracting the asymmetric trust-eroding effects of silence; maintaining recognizably consistent communicative personas across formal and informal digital channels; delivering all critical feedback through private rather than group channels, particularly in high power-distance cultural contexts where face threats carry amplified relational costs; modeling the organizational acceptability of uncertainty and error through proactive self-disclosure; and sharing organizationally significant information with employees before rather than after the formal communications process.

The findings also carry implications at the level of organizational system and process design, which should not be reduced to individual leadership behavior. These are interpretive implications derived from employees’ accounts rather than tested system-design interventions; they are offered as design considerations whose effectiveness remains to be evaluated in implementation research. Several of the negative interpretations documented here arise from conditions that organizations can engineer against rather than merely coach against. First, structured communication routines can institutionalize the acknowledgment behaviors that participants read as respect: Explicit organizational response-time norms (including a norm of brief acknowledgment when a substantive response is not yet possible) convert an individual supervisory virtue into a designed system property, reducing the incidence of the diagnostically damaging silences documented in Theme 1. Second, transparent decision records and proactive information-sharing workflows—for example, routinely published summaries of decisions affecting remote staff—directly target the informational asymmetry identified in Section 4.2.2, so that employees’ access to organizationally significant information does not depend on individual supervisory disposition. Third, coordination and workflow platforms shape the visibility conditions under which micro-behaviors are read: Integrated collaboration environments with shared task status reduce employees’ reliance on inferring supervisory attention from latency and presence indicators, whereas fragmented tool ecosystems amplify that reliance. Fourth, channel architecture can be designed to protect face: designating private, low-visibility channels as the default route for critical feedback embeds the face-saving practice documented in Section 4.3.1 into system configuration rather than leaving it to supervisory judgment. In engineering management terms, the interpretive vulnerabilities documented in this study suggest candidate requirements for the design of communication systems and management processes, not only for leadership training, subject to evaluation in the organizational contexts where they would be deployed.

For human resource (HR) professionals and leadership development programs, the findings challenge the assumption that leaders who perform effectively in co-located settings will automatically develop the digital communication competencies required for remote leadership effectiveness. Leadership development programs in Indonesian organizations and more broadly in high power-distance organizational contexts should include specific modules addressing the following areas: digital micro-behavior management and the systematic gap between intended and received signals; the cultural dynamics of face-saving in virtual feedback delivery; the permission-structure mechanism and practical strategies for credibly accelerating it; the asymmetric dynamics of trust accumulation and the disproportionate cost of supervisory silence; and the informational asymmetry of remote settings and the organizational equity implications of how that asymmetry is managed.

For multinational organizations operating in high power-distance Southeast Asian contexts, the findings suggest that leadership development frameworks designed in low power-distance Western organizational cultures cannot be directly transposed. The behavioral competencies that constitute effective remote leadership are culturally mediated: What constitutes an appropriate and effective trust-building or permission-granting signal in an Indonesian context may differ substantially from what employees in European or North American contexts read as similarly consequential. Cross-cultural leadership development should be designed with this interpretive specificity in mind.

For SDG alignment [43], the findings carry direct relevance. SDG 8’s call for productive employment and decent work for all is supported by identifying the specific behavioral conditions—consistency, proactive disclosure, culturally responsive feedback practices, deliberate digital presence management—that enable employees in remote and hybrid settings to exercise meaningful voice and engage fully with their organizational roles. SDG 10’s commitment to reducing inequalities within nations is addressed by surfacing the permission-structure mechanism as a systematic barrier to organizational participation for employees in high power-distance contexts—a barrier that constitutes a form of structural inequality in access to voice, and that specific leadership practices may help mitigate. Organizations and policymakers in Indonesia and comparable Southeast Asian contexts should treat inclusive remote leadership development not as a management efficiency concern but as an equity and decent work imperative. These implications are also consistent with evidence that sustainable leadership in Malaysian higher education depends on context-sensitive implementation, institutional integration, and adaptive leadership practices [44]. Related science-mapping research identifies transformational leadership and knowledge sharing as a growing domain, reinforcing the importance of examining how leadership behaviors shape information exchange and participatory outcomes in digitally mediated workplaces [45].

Several limitations require candid acknowledgment. The interpretive design prioritizes analytic depth over distributional representativeness; findings reflect the accounts of 48 employees in specific urban, knowledge-sector, and professional employment contexts, and should not be extrapolated to Indonesian remote workers more broadly without further inquiry. The sample is substantially weighted toward urban, educated, and professional-services-adjacent employment; rural, manufacturing, healthcare, and agricultural sector experiences, which represent the majority of Indonesian employment are entirely absent. This represents a significant boundary condition on the study’s claims, not merely a qualification.

The study relies exclusively on employee self-reports gathered through retrospective interviews, introducing possibilities of recall bias, retrospective sense-making, and social desirability effects. Supervisors’ perspectives are entirely absent, meaning the study captures employee-side perception of leadership rather than the dyadic, mutually constituted construction of leadership meaning. The gap between employee perception and supervisory intent, which this study argues is analytically central cannot itself be examined from a single-perspective design. A related interpretive boundary must be maintained: The findings warrant claims about how leadership behaviors are interpreted by employees and how those interpretations shape workplace responses, not claims about what leaders actually do or intend. A delayed response that an employee reads as inattention may in fact reflect workload, organizational constraint, or system limitations; the contribution of this study lies in explaining the interpretive consequences of such behaviors, not their causes, and conclusions about remote leadership practice should be read within that boundary. A further limitation concerns technological infrastructure. Digital micro-behaviors do not occur in a platform vacuum: A delayed response in an informal messaging application may carry a different meaning from the same delay in a formal project management environment, and organizations with integrated collaboration systems produce different visibility and communication patterns than those relying on disconnected tools. This study did not systematically analyze the platform and infrastructure environments within which participants’ interpretations were formed, and the relationship between digital infrastructure and leadership interpretation is accordingly identified as an important direction for future research rather than an established feature of the present account. The cross-sectional design means that the temporal dynamics of trust accumulation and permission-structure development, which participants described as unfolding over months, cannot be directly observed. Longitudinal interpretive claims should be treated as hypotheses for future investigation rather than established findings. The methodological tension between the study’s inductive orientation and its theoretically organized interview structure (Section 3.3) is a recognized limitation: Despite design features intended to preserve inductive discovery, the possibility of theoretical pre-sensitization cannot be fully excluded.

Negative case analysis identified six participants whose accounts qualified the dominant thematic patterns. P01 and P11 (both technology sector) described hierarchical norms as having substantially weakened in their remote work experience; P06 (Technology, Sumatra) described digital micro-behaviors as largely irrelevant to his trust assessments; P14 (Professional Services) described behavioral inconsistency as a trust signal rather than a trust barrier; and P02 and P09 described preferences for directive hierarchical clarity rather than participative openness. These accounts collectively caution against reading the study’s themes as universally applicable even within the sample. The themes describe dominant patterns across 36–48 participants; the negative cases define their analytical boundaries and scope conditions.

This study opens at least ten productive research directions: (1) Longitudinal qualitative designs tracking trust accumulation and permission-structure development in real time rather than through retrospective accounts. (2) Dyadic qualitative designs incorporating both employee and supervisor accounts to examine the gap between intended and received leadership signals—the asymmetry this study identifies but cannot directly measure. (3) Comparative qualitative studies examining whether the permission-accumulation process operates in similar form in other high power-distance Southeast Asian contexts. (4) Qualitative studies examining digital micro-behavior interpretation in low power-distance contexts to assess cultural specificity of the signaling frameworks identified here. (5) Mixed-methods designs using qualitative findings from this study to develop and validate behavioral measurement instruments for digital communication competence and permission-structure dynamics. (6) Research extending the sample to rural, manufacturing, healthcare, and agricultural sector employees in Indonesian organizational contexts entirely absent from the current study. (7) Experimental research testing whether interventions targeting the specific behavioral competencies identified here produce measurable improvements in psychological safety and voice in Indonesian remote teams. (8) Research examining the role of peer behavior, rather than supervisory behavior alone, in constructing the social proof basis for permission structures. (9) Investigation of gender dynamics in remote leadership perception in high power-distance Indonesian contexts. (10) Systematic examination of sector-level variation in permission-structure dynamics, given that government sector participants described markedly more rigid hierarchical constraints than technology sector participants. (11) Comparative investigation of how digital platform architecture and the integration of organizational collaboration infrastructure condition the interpretation of micro-behavioral signals, connecting the interpretive account developed here to the design of communication and workflow systems in engineering and technology-intensive organizations.

This inductive qualitative study set out to examine how employees in Indonesian organizations perceive leadership in remote work settings and what meanings they construct from those perceptions in relation to their sense of safety and willingness to engage. The answer that emerged from 48 in-depth interviews, analyzed through RTA, is both empirically grounded and theoretically consequential: Remote leadership is experienced not through formal style categories but through the cumulative texture of behavioral signals transmitted across digital channels. In high power-distance cultural contexts, those signals must be read, verified, and re-verified before the possibility of organizational voice can be genuinely accessed.

Four themes articulate this account. Behavioral consistency—the sustained alignment between what supervisors say and what they do across interactions, channels, and time—functions as the primary trust-generating mechanism in remote settings, operating through an asymmetric accumulation process in which positive gains accrue slowly and negative events, particularly silence, erode them rapidly. Supervisory disclosure of uncertainty, error, and organizational information carries heightened relational weight in digitally mediated settings, where the informational asymmetry between leaders and employees is structurally greater than in co-located environments, and where proactive sharing signals a relational orientation that converts compliance into engagement. The cultural renegotiation of hierarchy in Indonesian organizational contexts is not a passive adjustment to remote work conditions but an active, permission-dependent process in which employees must accumulate specific behavioral evidence that hierarchical norms have been genuinely overridden before voice becomes possible. And the communicative texture of digital presence—the micro-behaviors of camera usage, response latency, channel selection, and status management—constitutes a continuous, largely unintentional leadership signaling domain through which supervisory character, attentional regard, and authentic relational investment are continuously assessed.

An integrated process model of remote leadership interpretation emerges from this account. Digital micro-behaviors constitute the signal input through which employees infer supervisory character. Asymmetric trust accumulation and cultural permission structures then operate as partly parallel and interacting mechanisms: The former concerns perceived reliability and benevolence, while the latter concerns whether upward voice is culturally and organizationally legitimate. Together, these mechanisms condition whether psychological safety becomes available and whether employees engage in voice. This model extends knowledge-based trust theory and negativity-bias findings into remote leadership perception, identifies a signaling domain not captured by conventional communication-quality or LMX accounts, and provides a process-level explanation for how power-distance dynamics shape voice.

The practical implications for leadership development, HR policy, organizational design, and SDG-aligned decent work agendas in Indonesian and comparable Southeast Asian contexts are grounded in employees’ accounts and are offered as actionable starting points. Leadership that remote employees experience as effective is, in participants’ accounts, not reducible to digital skill or platform competence; it involves cultural attunement, behavioral consistency, and deliberate management of the signaling environment that employees inhabit. Organizations that treat remote leadership development as a purely technical challenge rather than a relational and cultural one risk systematically underinvesting in the capabilities that appear to matter most to the employees they seek to lead.

Future research should pursue the longitudinal, comparative, dyadic, and sector-diversified designs that this study’s necessarily bounded scope cannot provide. The conceptual model advanced here from remote leadership behaviors through signal interpretation, trust accumulation, and cultural permission structures to psychological safety and voice is offered as a theoretically grounded and empirically motivated proposition for investigation, not as a conclusion. The findings reported here are consequential; they are also a beginning.