Authentic leadership – for better and for worse? Leader well-being and inconsistency as moderating factors in the relation between daily authentic leadership and follower well-being

ABSTRACT

Despite comprehensive insights on favourable outcomes of authentic leadership for followers, there is a limited understanding about boundary conditions of authentic leadership. Leaders’ decreased well-being (i.e. high trait emotional irritation and low trait work engagement) or inconsistency in authentic leadership may attenuate the relation between authentic leadership and follower well-being (i.e. emotional irritation and work engagement). We tested these assumptions using a daily diary study design over five consecutive working days with a sample of 64 leaders and 162 followers. We found partial support for the cross-level moderation of authentic leadership inconsistency across these five days, as daily authentic leadership was related to daily work engagement (but not emotional irritation) of followers, when authentic leadership inconsistency was low. Contrary to our predictions, neither high leader trait emotional irritation nor low leader trait work engagement attenuated the relation between daily authentic leadership and daily follower well-being. The results point out to the importance of consistency in authentic leadership behaviour. However, leaders do not need to be overly concerned about negative consequences for followers when they lead authentically while having a reduced trait well-being.

KEYWORDS:

• Authentic leadership
• follower well-being
• leader well-being
• leadership consistency
• diary study

Authentic leadership is widely considered as a beneficial leadership style, which is related to positive outcomes for followers, organizations and leaders themselves (for an overview, see Gardner et al., 2011). Perceiving the authentic leadership in daily interactions (i.e., acting reflected and transparent; Walumbwa et al., 2008), can be an important resource for followers and may relate positively to their well-being. However, reflecting on criticism on authentic leadership and its conceptualization (e.g., Alvesson & Einola, 2019; Gardner et al., 2021; Helmuth et al., 2023; Ostermeier et al., 2023) as well as empirical insights on inconsistent leader behaviour (e.g., Kelemen et al., 2020), there may be boundary conditions to this relationship. We draw on uncertainty management theory (Lind & van den Bos, 2002) to propose that there may be two boundary conditions that may undermine the positive association between authentic leadership and follower well-being: First, decreased leader well-being and second, inconsistency in authentic leadership behaviour.

With our research, we aim to contribute to the debate between “advocats” and “sceptics” (Alvesson & Einola, 2019) by seeking a more nuanced understanding of authentic leadership. It has been argued that authentic leadership “is often unwanted” as “most people … have less sympathetic “authentic” orientations, for example, bad temper” and that in such case authentic leadership can sometimes raise social problems (Alvesson & Einola, 2019, p. 391). Those aspects can contradict the image of leaders as being reliable by providing security and certainty (Lind & van den Bos, 2002). Additionally, considering past findings about crossover effects of leader’s decreased well-being on their followers (e.g., Huang et al., 2016), this may be an important boundary condition for authentic leadership. This also pays into the debate by Helmuth et al. (2023) who question whether authentic leadership should be viewed solely in terms of positive psychology (Walumbwa et al., 2008), as the underlying concept of authenticity is neither positive nor negative (Kernis & Goldman, 2006). Thus, authentic leadership includes exhibiting and communicating both positive and negative well-being aspects (e.g., by expressing the leader’s true feelings or thoughts, or openly reflecting on the leader’s own weaknesses). Hence, we will investigate leaders’ well-being as boundary condition and examine whether authentic leadership is beneficial for followers’ well-being, even when leaders have a decreased trait well-being. We focused on trait leader well-being, as the influence of stable characteristics unfolds across different situations and over time. Hence, trait leader well-being can have more comprehensive consequences, as it is also strongly related to day-level expressions of their well-being across several days (e.g., Tims et al., 2011).

Furthermore, diary studies showed that leadership behaviour is highly variable (Kelemen et al., 2020) and can exhibit unsystematic (i.e., inconsistent) changes over time (McClean et al., 2019; Poethke et al., 2022). Hence, we propose that authentic leadership behaviour can fluctuate similarly, as leading authentically or not may depend on the context (Helmuth et al., 2023). When leaders are confronted with different people, situations, contexts, and demands, the degree to which leaders show high self-awareness and critical self-reflection can vary, e.g., for situations they are less familiar with (cf. McClean et al., 2019). Further, authentic leadership behaviours (e.g., relational transparency) may evoke different reactions in leader-follower interactions (cf. DeRue & Ashford, 2010), which might make it more likely for leaders to lead authentic in some situations than in others. Finally, acting out authentic leadership can require resources, which can again vary on a daily basis. However, acting according to one’s true or core self is essential to authentic leadership and fluctuations in those behaviours could be viewed as inauthentic and raise uncertainty in followers (cf. Lind & van den Bos, 2002). Thus, we aim to investigate perceived inconsistency in authentic leadership as a boundary condition. As we will investigate authentic leadership on a daily basis, we will be able to capture daily leadership behaviour and its relation with well-being on the same day. Since authentic leadership research has solely focused on between-person assessments of the leadership style, this study will allow more fine-grained insights into its within-person relations (Kelemen et al., 2020).

This study aims to make the following contributions. First, with our focus on boundary conditions, we can advance our understanding of when authentic leadership is related to follower well-being and can thus be better implemented in practice. Accordingly, we want to address whether leaders should still lead authentically even when they have a reduced trait well-being. It has been argued that when leader well-being is low, their authentic leadership can have negative consequences for followers (Alvesson & Einola, 2019). Furthermore, inconsistent behaviour contradicts the idea that authentic leadership emerges from a true or core self (Walumbwa et al., 2008), which in turn can limit its positive impact. Both factors, leader well-being and authentic leadership inconsistency are considered as unstable leader behaviours within uncertainty management theory (Lind & van den Bos, 2002) that links uncertainty to stress.

Second, we aim to get a better understanding of authentic leadership by having a look at authentic leadership on a daily basis. To our knowledge, this is the first diary study that assesses authentic leadership and well-being on a daily level. Accordingly, this multilevel approach allows to separate within- from between-level relationships, offering a nuanced view at which level authentic leadership relates to follower well-being (McCormick et al., 2020).

Finally, our results will also be of high practical relevance, as it could help leaders and practitioners to understand how authentic leadership is acted out in the best way. Though it can be relieving for leaders to be authentic about their reduced trait well-being (Weiss et al., 2018), it is important for leaders to know if this may be harmful to the well-being of their followers. Moreover, the results on the consistency in authentic leadership behaviour have important implications for how authentic leadership is trained and enacted in daily interactions.

Authentic leadership and follower well-being

Walumbwa et al. (2008) define authentic leadership as a multidimensional leadership style that is characterized by “greater self-awareness, an internalized moral perspective, balanced processing of information, and relational transparency” (p. 94). Authentic leaders have a sound understanding of themselves, e.g., knowing their own personality, emotions, abilities, strengths and weaknesses. Through the balanced processing of those aspects, they constantly deepen this knowledge as this means to be open to both, positive and negative self-related information. Further, through internalized moral perspective, authentic leaders show actions that are in line with their values and convictions. Relational transparency means to openly share and communicate with followers about one’s true emotions, opinions and self-related aspects. Authentic leadership is based on the multidimensional concept of authenticity by Kernis (2003). The dimensions self-awareness, balanced processing of information, and relational transparency of authentic leadership correspond to the dimensions awareness, unbiased processing, and relational authenticity of Kernis’ concept of authenticity (Helmuth et al., 2023). However, Walumbwa et al. (2008) added the fourth dimension of an internalized moral perspective. This deviates from Kernis’ work, in which authentic actions (i.e., individuals act in accordance with their true values) has been considered as the fourth dimension of authenticity. This shift in the conceptualization of authentic leadership towards a more moral and positive perspective has received criticisms (e.g., Alvesson & Einola, 2019; Helmuth et al., 2023). Despite this criticism, there is a significant conceptual overlap between authenticity and authentic leadership, which is mirrored in empirical work indicating a strong correlation between authentic leadership and authentic action (Helmuth et al., 2023).

There is meta-analytical evidence that authentic leadership is positively related to follower well-being (Zhang et al., 2022) and further empirical findings on the positive link between daily authentic leadership with positive emotions and authentic self-expression among followers (Yagil & Medler-Liraz, 2014). Hence, we propose that daily authentic leadership is positively related to daily follower well-being. Specifically, we focus on high work engagement as positive and low emotional irritation as negative well-being indicators. Work engagement is defined as “a positive work-related state of fulfilment that is characterized by vigour, dedication and absorption” (Schaufeli et al., 2006, p. 701). Vigour refers to a high level of energy and willingness to put persistently effort in the own work. Dedication is described as being strongly involved in the work while experiencing a sense of significance, enthusiasm, inspiration, and challenge. The final facet of work engagement, absorption, is characterized by being fully concentrated and engrossed in the work, with time passing by quickly (Bakker et al., 2023).

We argue that daily authentic leadership acts as a resource for individual followers (cf. Braun & Nieberle, 2017), which is positively related to their daily work engagement. In the daily work with their followers, authentic leaders exhibit a high relational transparency (i.e., clearly express their ideas and expectations) and signal their followers that they are truly interested in the followers’ ideas and points of view (e.g., Gardner et al., 2011). On such days, followers may feel motivated to put effort in their work, which is reflected in higher daily work engagement. As authentic leaders openly share their thoughts, emotions, or act according to their moral beliefs, they support their followers to reflect on their own motives and goals and, thus, engage themselves more in things that are meaningful to them which is reflected in higher work engagement (Bakker et al., 2023; Hannah et al., 2011; Lyubovnikova et al., 2017). Additionally, due to the leader’s clear direction (e.g., reflected in their self-awareness and internalized moral perspective), followers are more likely to experience higher work engagement, as goals and priorities are communicated and empathizes in a better way in their daily work (cf. Wei et al., 2016). These assumptions are also supported by previous studies, which found a positive link between authentic leadership and follower work engagement (e.g., Hsieh & Wang, 2015; Wei et al., 2016). Summarized, authentic leadership can make the followers’ work more engaging (Gardner et al., 2011).

As a negative well-being indicator, emotional irritation describes “subjectively perceived emotional strain in occupational contexts” that results from uncertainty (Mohr et al., 2006). Authentic leadership can be a meaningful daily resource against negative well-being (Ilies et al., 2005). Authentic leaders transparently provide information and share about their thoughts and emotions in their daily interactions with followers, which can reduce uncertainty and irritation (Lind & van den Bos, 2002; Mohr et al., 2006). Further, high self-reflection and self-awareness of authentic leadership can help followers to cope with work demands (Leroy et al., 2012), which again may be negatively related to daily irritation. Finally, authentic leaders are open to the expression of followers’ concerns and take care of their needs, buffering the followers’ experience of irritation at their daily work (cf. Laschinger & Fida, 2013). In sum, authentic leadership can be a preventive resource against impaired well-being such as emotional irritation (Gardner et al., 2011), which has been supported by empirical work (e.g., Braun & Peus, 2018; Laschinger & Fida, 2013; Rahimnia & Sharifirad, 2015). Thus, we hypothesize:

Hypothesis 1:

Daily authentic leadership behaviour will be negatively related to daily follower emotional irritation (H1a) and positively related to daily follower work engagement (H1b).

Authentic leadership: a matter of leaders’ well-being?

Authentic leadership depends on revealing personal aspects like opinions, thoughts or emotions, which also includes aspects of leader well-being (Walumbwa et al., 2008). Authentic leaders show and communicate to their followers how they feel and what impacts their well-being (Helmuth et al., 2023). Hence, leaders’ trait well-being should influence the relation between authentic leadership and possible follower consequences depending on whether leaders have a positive or negative trait well-being (cf. Braun & Nieberle, 2017). This assumption links authentic leadership to its theoretical basis, authenticity, which was not necessarily defined as a positive state or trait (Kernis & Goldman, 2006). Thus, it is of high relevance to consider if positive compared to negative self-aspects of the leader (e.g., in regard to the leader’s trait well-being) make a difference for their authentic leadership.

As explained above, authentic leadership should be positively associated with follower well-being (Gardner et al., 2011; Zhang et al., 2022). However, it is proposed that among leaders with reduced trait well-being or disengagement these positive relations may be mitigated or even reversed (Alvesson & Einola, 2019). As negative well-being aspect, we consider emotional irritation, as leaders with high levels of emotional irritation anger quickly and react irritable in interactions with others (Mohr et al., 2006). Although authentic leadership behaviour recognizes that leaders have negative self-aspects, reacting highly irritable in interactions with followers runs against the concept of authentic leadership, which can undermine possible beneficial consequences of authentic leadership for follower well-being (cf. Braun & Nieberle, 2017). Specifically, authentic leaders openly expressing about their emotional irritation on a daily basis can be – according to uncertainty management theory – a stressful experience for followers (Lind & van den Bos, 2002; Matta et al., 2017). This may undermine the positive consequences of internalized moral perspective, balanced processing, and self-awareness for leaders high in emotional irritation. In contrast, among leaders with low trait emotional irritation, their daily authentic leadership should be stronger related to follower daily well-being. Being less irritable when leaders talk about their own well-being or encourage followers to express their own points of view may be more beneficial for followers’ well-being.

While leading authentically, leaders’ high emotional irritation can raise uncertainty in daily interactions with followers in regard to adequate reactions. This is especially critical, as the leader-follower relationship is often hierarchical in its nature, followers might be uncertain about how to appropriately react to an open display of leader’s emotional irritation (e.g., when the leader asks for the followers’ critical perspective on an important issue in accordance with the balanced processing dimension of authentic leadership). Followers have to respond in an empathic and understanding way, yet they need to show respect at the same time and treat their leader in accordance with their more powerful position. In sum, we assume that leader’s decreased well-being (i.e., high trait emotional irritation) will moderate the relation between daily authentic leadership and follower well-being. Thus, the following cross-level interaction is hypothesized:

Hypothesis 2:

Leader emotional irritation will attenuate the positive relation between daily authentic leadership and follower well-being (H2a: low emotional irritation; H2b: high work engagement).

Well-being is more than the absence of illness (Topp et al., 2015). Hence, for a more comprehensive view, we have a look at the potential impact of positive leader well-being (i.e., trait work engagement) as well. Leaders with high trait work engagement show generally high levels of energy, involvement, and experience a sense of significance and enthusiasm (Schaufeli et al., 2006). This can strengthen the positive link between daily authentic leadership and follower well-being.

Authentic leaders have a sound understanding of their strengths and weaknesses (including their well-being aspects), which they transparently share with their followers. Hence, when they lead authentically and experience high work engagement, leaders may act as a role-model for their followers (Gardner et al., 2005). Consequently, work engagement in terms of the energy, involvement and enthusiasm embodied by the leaders matters for the consequences of their authentic leadership (Ilies et al., 2005). Since authentic leaders openly share their thoughts, attitudes and emotions (Gardner et al., 2011; Ilies et al., 2005), it is suggested that daily authentic leadership is especially related to follower well-being, when also the leaders have a high level of well-being like trait work engagement. Leaders with low trait work engagement, however, should share more negative aspects of disengagement, which may attenuate the positive link between authentic leadership and follower well-being (cf. Gutermann et al., 2017; Wirtz et al., 2017). Altogether, a cross-level interaction for leader work engagement is proposed:

Hypothesis 3:

Leader work engagement will strengthen the positive relation between daily authentic leadership and follower well-being (H3a: low emotional irritation; H3b: high work engagement).

Authentic leadership inconsistency

Previous research has pointed out the importance of inconsistency leaders show in their leadership behaviour. For example, Matta et al. (2017) extended the ideas of uncertainty management theory (Lind & van den Bos, 2002) and found that leaders being inconsistently fair had a more severe effect on physiological stress than being consistently unfair. Past work could also show that being consistent is especially important regarding authentic leadership (Peus et al., 2012), as authentic leadership is about expressing and acting in accordance with one’s “true” self (Walumbwa et al., 2008). If a leader expresses, communicates, and acts based on a specific conviction or value on one day and acts differently on another day, followers could get the impression that their leader is either not aware of his or her conviction or confirming to external situations or pressures (Gardner et al., 2005). Additionally, asking for the followers’ opposing opinion on one day and suppressing such efforts the next day leaves followers uncertain as to what the leader really wants (cf. Lind & van den Bos, 2002). Supposedly, the link between authentic leadership behaviour and well-being of followers can be attenuated or become even ineffective through this kind of inconsistency (Hsieh & Wang, 2015; Lind & van den Bos, 2002). Thus, the following cross-level interaction is proposed:

Hypothesis 4:

Authentic leadership inconsistency will attenuate the positive relation between daily authentic leadership and follower well-being (H4a: low emotional irritation; H4b: high work engagement).

Method

Sample and research design

This study utilized a daily diary design across five consecutive working days (i.e., from Monday to Friday) with a baseline questionnaire in the previous week, which is typical for diary studies in the leadership context (Kelemen et al., 2020). The study is part of a larger research project, which focuses on leadership, well-being of both leaders and followers and related processes. We recruited the six participating organizations via the research team’s professional network, newsletters, and social media. Organizations were from a variety of branches, including industry, the public and financial sector as well as consultancy. After an organization agreed to participate, the individual leaders were contacted and were able to decide, if they wanted to participate with their team. Participating leaders received study information and the link to the baseline questionnaire, which they forwarded to their team members. Leaders and team members were informed about the content and procedure of the study, that participation was voluntary, and that their organization would get feedback about the results of the overall organization. As an incentive, organizations were offered a presentation of the study results for leaders and followers and a 3-hour workshop for leaders.

Participants had one week to participate in the baseline study. At the end of the baseline study, participants were asked to give their email address in order to receive the invitations to the daily assessments. Participants received those emails and a reminder every day and were asked to fill in the questionnaire at the end of each working day. To match the participants’ responses across days and to their teams, the participants generated a pseudonymized code.

Before the study was conducted, approval was given by the ethical committee of the research institution of the authors as well as consent from the data protection officer (no 040_2020). Additionally, we recorded all hypotheses and related statistical analyses on the AsPredicted platform (https://aspredicted.org/blind.php?x=CFF_74D).

Data were collected between May 2021 and June 2022. Initially, 87 team leaders and 268 followers from 87 teams agreed to participate in our study. We excluded followers participating only on one day. After matching participant responses and exclusion of followers with only one daily measure, we had a final sample of 161 followers with 651 person-days and 62 leaders from 75 teams. On average, followers were 42.45 years old (SD = 11.29), mostly male (57%), and have been working for their organization for 8.65 years on average (SD = 9.66). Due to a technical error, this data is missing for about 50% of followers. The leaders were on average 46.25 years old (SD = 7.92), mostly male (70%), and have been working for their organization for 14.84 years on average (SD = 10.64). They had an average leadership experience of 10.28 years (SD = 7.30 years). On average, followers reported a daily contact time of 49.6 minutes (SD = 77.80 minutes) with their leaders.

Measures

All items were assessed with on a 7-point scale, ranging from 1 “Don’t agree at all” to 7 “completely agree”. Leaders rated their trait emotional irritation and trait work engagement in the baseline survey. Followers rated day-level authentic leadership, their day-level work engagement and emotional irritation in the daily survey. All daily measures were adapted to daily measurement whereby the items were reframed to match a time-frame for a specific day, which is common in diary research (Ohly et al., 2010).

Authentic leadership

Authentic leadership was assessed with one item from each of the four sub-dimensions of the Authentic Leadership Inventory (Neider & Schriesheim, 2011) in a German translation (Franke-Bartholdt et al., 2018; Reineboth et al., 2020). The items were selected based on a rating by four researchers in regard to adaptability for a diary study, e.g., to which degree items are observable in daily working life (cf. Ohly et al., 2010). All adapted items are presented in the online supplement. One example item is “My leader encouraged others today to voice opposing points of view.” To account for the multilevel nature of our data, we provide reliability estimates for each level of analysis and, thus, calculated McDonald’s Omega values based on a multilevel confirmatory factor analysis (see below; Geldhof et al., 2014). Reliability was ω = .75 at Level 1, ω = .91 at Level 2, and ω = .90 at Level 3.

Authentic leadership inconsistency

We operationalized authentic leadership inconsistency as the standard deviation across the daily authentic leadership measures for each follower. This operationalization has been used in previous research on leadership inconsistency (e.g., Johnson et al., 2012; Matta et al., 2017; Poethke et al., 2022). It represents the unsystematic changes in authentic leadership as deviations from the average leader’s authentic leadership style perceived by the follower. These deviations are spontaneous, difficult for followers to predict and understand, and, hence, may best represent the nature of the focal construct of authentic leadership inconsistency compared to other operationalizations of dynamism in leadership behaviour (see McClean et al., 2019; Poethke et al., 2022).

Emotional irritation

Daily emotional irritation of followers was measured with four items of the irritation scale by Mohr et al. (2005). Items were adapted to the daily measurement (e.g., “I am angry right now.”). One item of the original scale needed to be excluded, as it could not be adapted to the daily measurement. Reliability was ω = .82 at Level 1, ω = .94 at Level 2, and ω = .84 at Level 3. Leader trait emotional irritation was assessed with five items of the same scale (e.g., “I anger quickly”). Reliability was ω = .90 at Level 3.

Work engagement

In order to measure daily follower work engagement, a German version of the 3-item version of the Utrecht Work Engagement Scale (UWES-3; Schaufeli et al., 2006) was adapted to daily measurement (e.g., with the item “I am enthusiastic about my job right now.”). Reliability was ω = .83 at Level 1, ω = .94 at Level 2, and ω = .81 at Level 3. Leader trait work engagement was also assessed with the UWES-3 (e.g., with the item “I am enthusiastic about my job.”). Reliability was ω = .88 at Level 3.

Statistical analysis strategy

We had a three-level data structure of daily follower ratings (day-level or Level 1) nested in followers (follower-level or Level 2) nested in leaders (leader-level or Level 3). We used Bayesian multilevel modelling with uninformative priors to test our hypotheses. A Bayesian approach has several advantages, as it allows a more detailed evaluation of the model fit (e.g., convergence for every parameter) and is more suitable for small samples. Additionally, a Bayesian estimation is independent of p-values for hypothesis testing as it considers the posterior distribution of a parameter which provides the complete information about the probability of the possible parameter values (e.g., Kruschke & Liddell, 2018; Zyphur & Oswald, 2015). Particularly for estimating cross-level interactions, Bayesian estimation proved to provide accurate and unbiased results even for small samples (e.g., Stegmueller, 2013).

We followed recommendations about centring predictor variables in three-level models and applied cluster-mean centring at Level 1 and 2 (see Brincks et al., 2017). Hence, the daily follower ratings (Level 1) were centred on the respective follower mean. Additionally, we aggregated the daily follower ratings to the follower level. As multiple followers were nested in leaders, these aggregated ratings on the follower level (i.e., for authentic leadership) were themselves centred on the Level 3 mean. Furthermore, the aggregated follower ratings were also aggregated to the leader level. Aggregation to the follower and leader level can be justified by sufficient ICC1 (see Table 1) and average r_wg(j) values at the follower (r_wg(j) = .74) and leader level (r_wg(j) = .67; Bliese (2000). Prior to the analysis, we z-transformed (using the STANDARDIZE command in Mplus) all variables except for the outcomes to ease the presentation of the cross-level interactions.

Table 1. Descriptive statistics and intercorrelations.

	ω	ICC1	ICC2	1.	2.	3.	4.	5.	6.
Day Level
1. Authentic Leadership	.75	-	-
2. Emotional Irritation	.82	-	-	.00
3. Work Engagement	.83	-	-	−.03	−.20**
Follower Level
1. Authentic Leadership	.91	.71	.88
2. Emotional Irritation	.94	.64	.84	−.25**
3. Work Engagement	.94	.76	.91	.31**	−.47**
4. Authentic Leadership Inconsistency	-	-	-	−.17	−.04	−.04
Leader Level
1. Authentic Leadership	.90	.40	.86
2. Emotional Irritation	.84	.24	.75	−.05
3. Work Engagement	.81	.30	.80	.42**	−.49**
4. Authentic Leadership Inconsistency	-	-	-	-	-	-
5. Leader Trait Emotional Irritation	.90	-	-	−.04	.27	−.26	-
6. Leader Trait Work Engagement	.88	-	-	−.12	.11	.19	-	.40**
M	-	-	-	4.52	1.68	4.38	0.80	2.46	4.81
SD	-	-	-	1.71	1.15	1.47	0.59	1.21	1.09

ω = McDonald’s omega calculated from multilevel confirmatory factor analysis model.

*p < .05,**p < .01.

To test our proposed cross-level interactions, we followed the recommendations of Aguinis et al. (2013) and modelled the relationships between daily authentic leadership and daily work engagement as well as daily emotional irritation as random slopes on the lowest level (i.e., day-level or Level 1). Furthermore, we entered authentic leadership inconsistency as cross-level moderator on the follower level (i.e., Level 2), and leader emotional irritation as well as leader work engagement as cross-level moderators on the leader level (i.e., Level 3). Hence, these moderators did cross-level influence the Level 1 random slopes of daily authentic leadership and daily work engagement as well as daily emotional irritation. The syntax for this model is presented in the online supplement.

For the Bayesian estimation, we used 500,000 Markov-Chain Monte Carlo (MCMC) iterations with a burn-in phase of the first 250,000 iterations. To evaluate Bayesian model fit and MCMC convergence, we considered the potential scale reduction (PSR) value and inspected trace plots for every parameter (Depaoli & van de Schoot, 2017).

Construct validity

To test factorial validity, we estimated a multilevel confirmatory factor analysis (MCFA) with a Bayesian estimation and weak informative priors. We followed the recommendations of Muthén and Asparouhov (2012) to specify our model and priors distributions. In particular, we included residual covariances among the items for each model. For the unstandardized factor loadings, we applied a normal-distributed prior of N(1, 0.1). The prior distribution of the residual variances for the items on each level has been specified as an inverse-Wishart distribution of IW(1, p +6). For the residual covariances, we used a small-variance prior of IW(0, p +6). The MCFA model has been estimated with 1,000,000 MCMC iterations, whereas the first 500,000 iterations served as burn-in phase. To reduce the degree of autocorrelation, we used only every 10th iteration (Depaoli & van de Schoot, 2017).

The model had three latent factors on Level 1 and 2 (i.e., authentic leadership, follower emotional irritation, and follower work engagement), and five latent factors on Level 3 (i.e., authentic leadership, follower emotional irritation, follower work engagement, leader emotional irritation, and leader work engagement). The Bayesian model fit was good (Posterior Predictive Checking using Chi² = [−80.45; 79.68], posterior-predictive p-value = .506) with a PSR under 1.01 after about 340,000 MCMC iterations and trace plots exhibiting a typical pattern of chain convergence. Thus, factorial validity was supported. Drawing on the results of the MCFA model, we calculated McDonald’s Omega to estimate the multilevel reliability of our measures (Geldhof et al., 2014).

Results

The descriptive statistics (including mean, standard deviation, intercorrelations, intraclass correlations, and reliability estimates) are shown in Table 1. All hypotheses have been tested in a single Bayesian multilevel model (see Table 2). We considered different criteria to evaluate Bayesian model fit and MCMC convergence. As some parameters revealed some degree of autocorrelations. Thus, we thinned the posterior distributions by using only every 10th MCMC iteration to reduce the degree of autocorrelation (Depaoli & van de Schoot, 2017). PSR fell below 1.01 after about 39,000 MCMC iterations, and the trace plots showed a typical pattern of MCMC convergence.

Table 2. Results from Bayesian three-level model.

		Emotional irritation		Work engagement
Level 1: Day Level		B	95% CI	B	95% CI
Authentic Leadership		.04	[−.08; .17]	.08	[−.06; .21]
Level 2: Follower Level
Authentic Leadership		−.21*	[−.38; −.04]	.30*	[.08; .51]
Authentic Leadership Inconsistency		−.09	[−.29; .12]	−.03	[−.28; .23]
Level 2 Crosslevel Interaction: AL x AL Inconsistency		−.01	[−.09; .08]	−.11*	[−.21; −.01]
Level 3: Leader level
Authentic Leadership		−.14	[−.37; .10]	.48*	[.18; .77]
Leader Trait Emotional Irritation		.07	[−.16; .29]	−.01	[−.31; .28]
Leader Trait Work Engagement		.08	[−.19; .33]	.18	[−.15; .51]
Level 3 Crosslevel Interaction: AL x Leader Emotional Irritation		.08	[−.10; .20]	.04	[−.11; .18]
Level 3 Crosslevel Interaction: AL x Leader Work Engagement		−.01	[−.11; .09]	.03	[−.10; .15]

Table contains unstandardized coefficients.

*95%-CI excludes zero.

Regarding Hypotheses 1a and 1b, daily authentic leadership was neither related to daily follower work engagement (B = .08, 95%-CI = [−.06; .21]), nor daily follower emotional irritation (B = .04, 95%-CI = [−.08; .17]), as the posterior distributions of these relationships included zero as plausible value. Thus, Hypotheses 1a and 1b were not supported. Contrary to our predictions for Hypotheses 2a and 2b, leader trait emotional irritation did not cross-level moderate the relation between daily authentic leadership and daily follower work engagement (B = .04, 95%-CI = [−.11; .18]) as well as daily follower emotional irritation (B = .08, 95%-CI = [−.10; .20]).

Similarly, we did not find a cross-level interaction of leader trait work engagement for the relation of daily authentic leadership and daily follower work engagement (B = .03, 95%-CI = [−.10; .15]) and, respectively, daily follower emotional irritation (B = −.01, 95%-CI = [−.11; .09]). Hence, Hypotheses 3a and 3b were also not supported. Finally, we found a cross-level interaction of authentic leadership inconsistency on the relation of daily authentic leadership and daily follower work engagement (B = −.11, 95%-CI = [−.22; −.01]). In support for Hypothesis 4b, higher levels of authentic leadership inconsistency attenuated this relationship, as daily authentic leadership was only positively related to daily follower work engagement, when authentic leadership inconsistency was low (see Figure 1). However, and in contrast to Hypothesis 4a, we could not find such a cross-level interaction for daily follower emotional irritation (B = −.01, 95%-CI = [−.09; .08]).

Figure 1. Cross-level interaction plot.

Besides these results, we also found that follower-level authentic leadership was related to average follower work engagement (B = .30, 95%-CI = [.08; .51]) and to average follower emotional irritation (B = −.21, 95%-CI = [−.38; −.04]). Finally, team-level authentic leadership was also positively associated with average team-level work engagement (B = .48, 95%-CI = [.18; .77]).

Supplementary analyses

We ran a supplementary analysis to control for the influence of daily contact time between followers and their leader by excluding those daily observations (n = 182) without direct contact and adding contact time as a control at the day level (see online supplement Table S2). The pattern of results for our hypotheses remained largely unchanged. The only difference concerned the cross-level interaction of authentic leadership inconsistency on the relation of daily authentic leadership and daily follower work engagement, which was not supported in this model (B = −.03, 95%-CI = [−.15; .08]). Besides this, daily contact time was positively related with daily irritation (B = .16, 95%-CI = [.02; .28]).

Discussion

In this diary study, we investigated the relationship between daily authentic leadership behaviour and daily follower well-being and whether this relationship is cross-level moderated by leader trait well-being and by authentic leadership inconsistency. The first hypotheses were not supported, as daily authentic leadership was neither related to daily follower work engagement nor to daily follower emotional irritation. Interestingly, though the relation did not find support on day level, authentic leadership was related to work engagement and emotional irritation on follower level and to work engagement on team level. Additionally, we only found support for the cross-level moderation of authentic leadership inconsistency, but not for leader well-being.

Theoretical implications

This is one of the first studies that offers insights into daily authentic leadership and its linkages with follower well-being (Kelemen et al., 2020). Our findings are a good example of the need to investigate possible differences between relationships at within- and between-levels as explicated by (McCormick et al., 2020), especially given the fact that they report substantial differences between levels for leadership behaviour. Contrary to our predictions, the suggested relationships between authentic leadership and follower well-being were not found at the within-person level in this study, which stands in contrast to the relations that were found at the follower and team level in this and also in previous studies (Zhang et al., 2022). This is especially interesting as previous diary studies have found relations with daily well-being for other leadership behaviours like transformational leadership (Kelemen et al., 2020).

These results may contribute to our understanding of authentic leadership: It may not make a difference for follower well-being whether a leader shows more authentic leadership behaviour on a specific day (i.e., within-person relations). Rather, the degree to which a leader shows authentic behaviour in general is related to follower well-being (i.e., between-person relations). Thus, authentic leadership may not directly translate into followers feeling better. Rather, the benefits from authentic leadership behaviour that were shown on the between-person level could be due to the overall perception of the leader as reliable and predictable (Lind & van den Bos, 2002), as well as the relationship between leader and follower (Gill & Caza, 2018; Yagil & Medler-Liraz, 2014). Hence, high levels of self-awareness and transparency that are shown on the between-person level by authentic leaders may translate into a positive leader-follower relationship and positive team climate, which is related to the well-being of followers. At the same time, it must be noted that the between-person results also have to be interpreted as within-team results, as the followers are nested in teams. Hence, the between-person results indicate that followers, perceiving more authentic leadership relative to their team colleagues, likely experience lower irritation and higher work engagement compared to their colleagues in the same team. Besides this, authentic leadership was also related to average follower well-being on the team level. Thus, teams as a whole may experience higher work engagement, when their leaders show generally high levels of authentic leadership.

While further studies are required to replicate the current findings, the results contribute to a promising debate on differentiated effects of leadership between team, follower, and day levels (Avolio & Gardner, 2005). For the successful implementation of leadership, it is crucial to know if and how leadership behaviours are related to well-being on the within- and between-person levels (Kelemen et al., 2020) as well as on the individual and on team level respectively.

In regard to the second hypothesis, the results may provide first evidence against the criticism that authentic leadership is not a suitable or practicable leadership style as it could lead to adverse reactions among followers and cause “social problems” (Alvesson & Einola, 2019, p. 391). Our results offer first insights that leading authentically while generally having a decreased trait well-being (i.e., high trait emotional irritation), does not necessarily have detrimental effects on follower well-being. To test the robustness of this finding, further research may focus on the impact of other leader traits like the dark triad of personality (Paulhus & Williams, 2002). At the same time, positive well-being (i.e., high trait work engagement) does not strengthen this relationship either, which challenges the idea of increased crossover and trickle-down effects of well-being (Braun & Peus, 2018; Dietz et al., 2020; Köppe et al., 2018) from leaders to followers through authentic leadership.

The results contradict our hypothesis, based on uncertainty management theory (Lind & van den Bos, 2002), that the interplay of authentic leadership with generally lower leader well-being would exacerbate follower stress. However, our results give only first insights that need to be replicated and elaborated on in future studies. Particularly, it remains unclear whether leader well-being does indeed not influence the relationship between authentic leaders and follower well-being, as (1) there could be positive and negative consequences of leading authentically with decreased well-being that balance each other out or (2) as leader well-being can influence follower well-being independent of leading authentically or not.

Regarding the first point, Helmuth et al. (2023) suggested that conceptualizing authentic leadership only as a positive or morally-based leadership style does not do justice to the construct. This is also emphasized, as the dimensions of authentic leadership differ from the conceptualization of the underlying construct of authenticity (Helmuth et al., 2023; Kernis, 2003). Hence, the concept of authentic leadership may be in need of a revision (Alvesson & Einola, 2023), as it could be framed in a neutral way and more strongly focused on observable (i.e., measurable) actions (Helmuth et al., 2023). This means that the leader is aware of, shares, and shows positive and negative aspects of the self. Such transparent behaviour of the leader may outweigh the costs of dealing with negative thoughts or attitudes that emerge from decreased well-being of the leader among the followers. This may encourage followers to openly share about their own weaknesses without having to fear negative consequences (cf. Edmondson, 1999).

Regarding the second point, it is possible that the well-being of leaders does impact the perception of and interaction with their followers, independent of leading authentically or not (Dietz et al., 2020; Matick et al., 2022). Both positive and negative aspects of the leaders’ trait well-being may affect their thoughts and behaviours. Even if they do not openly express and explain about it, followers can notice and experience the consequences of leader well-being. Yet, when leaders are authentic, it could be easier for followers to understand their behaviour. This might help followers to attribute and deal with their leaders’ behaviours accordingly and may lessen the negative effects of inconsistency since followers are better able to understand why leaders’ behaviour is different across different situations (Schilling et al., 2023).

Finally, hypothesis 4 found partial support as daily authentic leadership was positively related to daily follower work engagement when authentic leadership was consistent over time, while there was no such effect for daily follower emotional irritation. This result seems to be paradoxical at a first glance, as a day-level relation between authentic leadership and follower work engagement implies that a higher level of authentic leadership (i.e., a deviation from the average level of authentic leadership) is related to a higher level of work engagement, which somehow contradicts the assumption of consistency. At a second glance, however, these small deviations from the average level of authentic leadership may make a difference for the followers’ engagement, when the leader is otherwise consistent in his or her behaviour (Michel & LeBreton, 2011; Poethke et al., 2022).

This is an important finding, as being consistent in authentic leadership behaviour can be considered a precondition to be viewed as an authentic leader in the first place (Gardner et al., 2005). Authentic leadership behaviours on a specific day can then make a difference as they are not perceived as a deviation from the usual leader behaviours. The hypothesis was not supported for follower emotional irritation, which is in accordance with the other results, as the relations for authentic leadership were stronger for work engagement than for emotional irritation. This could imply that authentic leadership is rather a promoting factor for positive well-being than a buffer against decreased well-being (Macik-Frey et al., 2009). However, we could not replicate the cross-level interaction regarding work engagement in a supplementary analysis controlling for contact time and excluding daily observation without contact between follower and leader. Hence, this finding needs to be interpreted with caution.

Limitations and future directions

There are some limitations of this study that need to be addressed, offering also several approaches for future studies. First, the assessment was conducted at the end of each working day in order to allow the assessment of the leadership behaviour throughout the whole day and the assessment of follower well-being resulting from the working day. Though a diary study captures actual daily behaviour in a more specific way, it is still not a record of actual behaviours and interactions that would e.g., capture if authentic leaders actually expressed their well-being. Hence, future studies may use an experience sampling design to assess leader-follower interaction directly after they occurred (Ohly et al., 2010) or by developing more specific measures that assess the expression of specific emotions or well-being aspects. Further, though leadership and follower well-being were assessed daily, the results do not allow for causal interpretation (Antonakis et al., 2010). However, we aimed to establish the temporal precedence and direction of these relations, as we measured daily authentic leadership retrospectively for the working day and follower well-being with respect to the current state (i.e., at the end of the working day; see Spector, 2019). Future studies may engage in research designs that allow for the examination of reciprocal follower-leader processes (Rudolph et al., 2020).

It was surprising that daily authentic leadership was unrelated to daily follower irritation and work engagement. Besides theoretical reasons (see above), the high ICCs at the follower level – indicating that most variance could be accounted to this level – can explain the absence of these relationships. Regarding our unsupported cross-level moderation of leader well-being, the timeframe of our study could be of interest. Our daily diary design aimed at examining rather short-term changes (i.e., day-to-day). As a speculation, the attenuating effect of leader’s decreased well-being may rather unfold over weeks or months. Thus, future studies may use weekly-diary studies or longitudinal studies over a longer period to investigate the interplay of leader’s decreased well-being and authentic leadership. Additionally, we have used multi-source data (i.e., the leaders rated their own trait well-being, whereas the followers rated daily authentic leadership and their daily well-being). While this is a suitable strategy to reduce common method bias (Podsakoff et al., 2003), it could be possible that the followers were not aware of or did not perceive the leaders’ decreased well-being. Besides this, also omitted variables could have an influence on the interaction. It could be possible that leaders with high self-control capacity or emotion regulation could be better able to manage their high trait emotional irritation at work. Likewise, follower-related variables like their implicit leadership theories or situational variables may be of importance. This could also explain the lack of a cross-level interaction effect. Furthermore, our sample size on Level 3 is relatively small. Although Bayesian estimations are appropriate to test cross-level interactions with small samples (Stegmueller, 2013), future studies may try to replicate our findings regarding the interactions of leader well-being and authentic leadership with a larger sample to test their robustness.

Finally, it is suggested that leadership inconsistency needs to be differentiated. The variances in authentic leadership behaviour that were described in this study need to be distinguished from functional adaptations of leadership behaviour to the requirements of different kinds of situations and followers (Michel & LeBreton, 2011; Poethke et al., 2022). Such coherence includes fluctuations in leadership behaviour that are “consistent, reliable and predictable” and a “purposeful and systematic variability in a leader’s behaviour and decisions” (Michel & LeBreton, 2011, p. 688). Supposedly, it makes a difference to the perceptions of followers if leaders differ in their behaviour across comparable situations, compared to acting differently across different contexts and circumstances.

Insights on authentic leadership are especially interesting since authentic leaders need to adapt their behaviour to different situations and followers, yet being perceived as authentic still requires leaders to be consistent and act based on their core or “true” self. It would thus be worthwhile for future studies to investigate on the interplay of self- and follower-assessed authentic leadership, situational demands and the organizational culture (Avolio & Gardner, 2005; Munyon et al., 2021).

Practical implications

There are a number of practical implications that can be derived from this study. However, as future research is needed to replicate our findings, these implications have to be interpreted with caution. First, in line with previous findings (e.g., Johnson et al., 2012; Matta et al., 2017), the results may underline the importance of consistency in one’s leadership behaviour. Leaders need to be aware that showing behaviours that are perceived as inconsistent can have detrimental effects in regard to followers’ well-being, especially when it comes to acting inconsistent in regard to behaviour that reflects the “true self” of a leader. It can thus be helpful for leaders to reduce perceived inconsistency by communicating and explaining about the motives and criteria for their behaviour (Schilling et al., 2023).

Second, the results give first evidence that the relation between authentic leadership and follower well-being is not mainly due to daily behaviour on the within-person level, but on the between-person level. Authentic leaders do not need to share every single thought or emotion on a daily basis, but they need to constantly invest into an authentic leader-follower relationship, e.g., by letting their followers generally know about their attitudes and approaches. By this, authentic leadership can provide a valuable approach to followers’ well-being.

Third, the results give initial support to the idea that authentic leaders may do not need to be overly concerned about sharing even negative aspects of their well-being, while we still consider the way it is expressed as crucial. It might surprise some followers if leaders share about their lacking engagement for a specific job. Yet, at the same time, this provides an opportunity for followers to understand and identify with their leaders, and it is further a chance for leaders to act as a role model and show that one can face challenges and deal with less enjoyable work experiences.

Conclusion

This study investigated whether leader well-being and leader inconsistency are boundary conditions of authentic leadership that are related to decreased follower well-being due to increased uncertainty of followers. This idea was only supported for leadership inconsistency in regard to follower work engagement (but not to emotional irritation). This indicates that leaders do not need to be overly concerned in showing even negative well-being, but it underlines the importance of acting out authentic leadership consistently.

Disclosure statement

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

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/1359432X.2024.2361511