Authoritarian leadership styles and performance: a systematic literature review and research agenda

Although authoritarian leadership styles are often associated with negative performance, work climate deterioration, increased power distance, and centralized control, contradictory empirical evidence has emerged in the literature. In this paper, we perform a systematic literature review with three aims: (1) understand the effects of authoritarian leadership styles on performance, (2) study the temporal and geographical evolution of the scientific debate, and (3) establish a research agenda for the future. The results show that in the last two decades, the interest for the field has shifted from Western to Eastern countries. Moreover, many authors encourage leaders to increase or decrease their degree of authoritarian leadership depending on the context to more effectively connect leadership with performance. Therefore, leadership should be studied in light of a more complex approach that considers hybrid leadership styles and their effects on performance. Finally, we discuss our study’s limitations and managerial implications.


Introduction
Authoritarian leadership styles involve high levels of control over subordinates (Chiang et al. 2020). Authoritarian leaders tend to use their authority, which is ensured by organizational hierarchies, to demand absolute obedience of their 1 3 followers (de Hoogh et al. 2015). Superiors adopting these leadership styles tend to centralize their power and accentuate the power distance between them and their subordinates (Schaubroeck et al. 2017). Evidence in the literature has shown that authoritarian leaders press their subordinates to achieve demanding objectives and to follow the rules (Li et al. 2018;Karakitapoğlu-Aygün et al. 2021). Although at first glance, these leadership styles may not seem to ensure a good work climate or facilitate high performance (Shen et al. 2019), the empirical literature has shown contrasting results during the last decades.
Authoritarian leadership styles are often associated with negative performance, complex leader-follower relationships, and high intentions of followers to leave (Schaubroeck et al. 2017). For example, Chiang et al. (2020) showed that the work climate worsens if authoritarian leaders operate by suppressing subordinates' emotions. Schuh et al. (2012) showed that subordinates' effort could be limited if superiors act as authoritarian leaders. Schaubroeck et al. (2017) demonstrated the direct relationship between subordinates' disapproval of power distance and adverse effects of directive leadership on performance. Therefore, authoritarian leadership styles seem to clash with the high dynamism of the new globalized and hyperconnected markets.
Nevertheless, several studies have identified specific conditions under which authoritarian leadership styles can positively affect workgroup performance. For example, positive results can be achieved in workgroups characterized by high levels of traditionality and guided by authoritarian leaders (Shen et al. 2019). Directive leadership can ensure good outcomes when rewards are low, group size is large, and failure is not too costly (Rahmani et al. 2018). Moreover, authoritarian leaders can succeed in workgroups with low team power struggles (de Hoogh et al. 2015) and high participation (Sagie 1996). Finally, Karakitapoğlu-Aygün et al. (2021) highlighted that authoritarian leadership styles can positively affect communication if authority is not exaggerated. These contrasting results highlight the need for a comprehensive review of the evolution of the scientific debate on this topic.
Furthermore, cultural prejudice about the effectiveness of authoritarian leadership styles should be addressed in light of the complexity of new businesses. For example, authoritarian and paternalistic leadership styles are "still predominant in many Asian cultures" (Shen et al. 2019: 498), where Confucianism is widespread and applied in business (Shen et al. 2019). In particular, the literature highlights the "necessity of challenging the deeply rooted beliefs held by many Chinese managers that authoritarian leadership is an effective leadership strategy" (Li et al. 2019: 951). Nevertheless, authoritarian leadership styles are popular in contemporary business organizations worldwide (Chiang et al. 2020). Therefore, empirical evidence reveals a contradiction between organizational theory and practice: even though authoritarian leadership styles are formally considered ineffective, management employs this kind of leadership in practice worldwide. The literature highlights the importance of "leadership styles prevalent in the culture under investigation and examines its cultural roots and dominant psychological mechanisms" (Chen et al. 2014: 813). Consequently, questions arise about the evolution of the authoritarian leadership concept and applications of authority in leaders' behaviors.

3
Authoritarian leadership styles and performance: a systematic… A considerable number of literature reviews have debated the effects of leadership styles on performance. For example, Yahaya and Ebrahim's (2016) review on this topic was limited to transformational, transactional, and laissez-faire leadership styles. Harrison et al. (2016) devoted their systematic review to entrepreneurial leadership and its effects on performance. Georgakakis et al. (2019) analyzed the role assumptions of CEOs and top management teams, organizing them into theoretical categories. However, their research did not classify authoritarian leadership styles depending on these assumptions. Laureani and Antony's (2017) comprehensive literature review discussed the effects of a general conceptualization of leadership and its effects on Lean Six Sigma. Servant leadership's effects on performance were debated by Langhof and Güldenberg (2019) and Parris and Peachey (2013). The general effects of leadership on performance were discussed by Asrar-ul-Haq and Anwar (2018). To summarize, although the literature highlights the importance of leadership for performance, we could not find a review devoted to authoritarian leadership styles and their outcomes.
For these reasons, we concentrated on authoritarian leadership styles-i.e., authoritarian, autocratic, directive, and paternalistic leadership -to explore literature findings of their effects on performance. In particular, we perform a systematic literature review (SLR) to answer the following research questions: RQ1: What are the effects of authoritarian leadership styles on performance? RQ2: What has been the temporal and geographical evolution of the scientific debate concerning the effects of authoritarian leadership styles on performance? RQ3: How does the discussion about authoritarian leadership styles make sense in light of the strong dynamism of new markets?
The article is structured as follows. We define the key concepts, i.e., authoritarian, autocratic, directive, and paternalistic leadership. After that, we describe in detail the methodology employed for the selection and analysis of the literature. The subsequent descriptive analysis shows the results of publications' years, typologies, and fields and the evolution of authors' and statistical units' origins over time. Then, we develop the content analysis with a conceptual map of the field, an analysis of the most employed theories, the themes that emerged, and the future research opportunities identified by the authors of selected articles. Finally, we conclude our article with a general discussion of the results and indications for the future expansion of this study.

Definitions of key concepts
In this study, we refer to four leadership styles: authoritarian, autocratic, directive, and paternalistic. In this section, we present four definitions derived from the extracted articles' content in our dataset. In the past, the literature treated authoritarian, autocratic, and directive leadership styles interchangeably (Chiang et al. 2020). However, commonalities and differences have emerged during the most recent debate.
Authoritarian leadership styles "include exercising discipline, authority, and control over followers" (Karakitapoğlu-Aygün et al. 2021: 475). They demand that employees meet high work standards and reprimand employees for poor performance (Wang et al. 2013;Lee et al. 2019). To achieve these goals, authoritarian leaders exhibit high self-confidence and plan their actions to ensure that their subordinates do not challenge their authority. Authoritarian, autocratic, and directive leaders limit "followers' autonomy and self-determination, whereby leaders control followers via impersonal procedures and rules" (Li et al. 2019: 931). They provide "clear directions and expectations regarding compliance with instructions" (Sanchez-Manzanares et al. 2020: 840). After that, they tend to centralize decisions and limit subordinates' opportunities to express their opinions (Yun et al. 2005).
In addition to revealing commonalities among authoritarian leadership styles, the literature shows substantial differences. According to Chiang et al. (2020), authoritarian and autocratic leadership styles differ in two main characteristics. First, evidence in the literature does not associate autocratic with destructive leadership styles, while authoritarian leadership is often associated with the 'dark side' of leadership. Second, autocratic leaders are task-oriented and therefore are accepted by subordinates. Authoritarian leaders, instead, trigger feelings such as fear of distrust. Moreover, authoritarian and directive leaders differ in the way they give subordinates feedback. In particular, "a directive leader focuses on providing guidance", while an authoritarian leader "focuses on controlling and making demands of subordinates" (Chiang et al. 2020(Chiang et al. : 1085. Moreover, authoritarian behaviors are part of the construct of paternalistic leadership, which "combines strong discipline and authority with fatherly benevolence and moral integrity in a personalistic atmosphere" (Wu et al. 2012: 97). Therefore, paternalistic leaders have a genuine interest in subordinates' well-being in both their professional and private lives (Hiller et al. 2019). The construct of paternalistic leadership consists of three dimensions: authoritarianism, benevolence, and morality (Chen et al. 2014;Chou et al. 2015;Hiller et al. 2019). The most crucial difference between paternalistic leaders and authoritarian, autocratic, and directive leaders is the benevolent side of this style. While morality could be associated with directive leadership and authority could be ascribed to authoritarian and autocratic styles, benevolence can be ascribed only to paternalistic behaviors.

Methodology
SLR is a methodology characterized by a rigorous protocol (Denyer and Tranfield 2009;Post et al. 2020), in which authors' interpretation and creativity are limited to achieve the highest possible level of objectivity (Boell and Cecez-Kecmanovic 2015;Snyder 2019). According to most authors who have debated the SLR methodology, the SLR procedure is a fit when a researcher seeks proof in the literature to answer specific research questions. Our research examines a small group of leadership styles, i.e., authoritarian styles, and their effects on firm and employee performance. Given this level of specificity, the SLR was the best available methodology for answering our questions.

3
Authoritarian leadership styles and performance: a systematic… The protocol we chose for our SLR was proposed by Wolfswinkel et al. (2013), who introduced the application of grounded theory (GT) for developing content analyses in SLRs. GT (Corbin and Strauss 1990;Strauss and Corbin 1997) is a research method based on the interpretative paradigm about sociological research, which interprets the processes underlying a specific phenomenon. Using open, axial, and selective coding (Table 1), the research starts from a set of documents and organizes their contents into subthemes, themes, and, eventually, superordinate structures.
The advantage of using GT is that researchers concentrate on concepts and theories to reveal themes debated in the literature during the analysis; the themes are not decided before the study (Wolfswinkel et al. 2013). Given that our research was not driven by any previous convictions about the effects of authoritarian leadership styles on performance and was not influenced by cultural prejudice, GT was the best possible choice to develop our content analysis. Table 1 shows the detailed steps of the protocol we followed, and Fig. 1 shows a PRISMA 2020 (Page et al. 2021) flowchart that displays the process of the extraction and selection of relevant articles. Finally, from 2011 to 2021, the average was 3.85. Therefore, the increasing number of publications and authors involved in the debate demonstrates the increasing interest in the topic. In particular, the distribution of articles per year shows a considerable focus on empirical research. In the dataset, 47 articles are empirical, and only seven are theoretical.

Publications' years, typologies, authors, and fields
Our research covered a wide range of journals and fields. The fields interested in the topic also changed over time, as shown in Fig. 3a and b. From 1966 to 2010, 48% of the articles were published in organizational psychology journals. In the next decade, this percentage decreased to 29%. Organization studies journals published 19% of the articles from 1966-2010, while from 2011 to 2021, this percentage increased to 22%. From 1966 to 2010, there were two articles published in general psychology journals, and in the next decade, there were none. General management, ethics, and social responsibility journals published 11% of the articles from 1966 to 2010 and 18% from 2011 to 2021. Finally, a new considerable area emerged during the last decade of analysis: international business and area studies journals published 15% of the articles from 2011 to 2021.
To summarize, during the analyzed period, the topic was debated in journals in the fields of psychology, organization studies, general management, ethics, and We used the SCOPUS and Web of Science electronic databases for three fundamental reasons. The first is the considerable number of journals included in their archives. Second, they are frequently used for developing systematic literature reviews (e.g., de Freitas and Costa 2017; Liao et al. 2017). Third, their frontends are user-friendly and allow the export of many publications' data 1.4 Decide on the specific search terms After several exploratory investigations and after reviewers provided suggestions, we searched the title, abstract, and keywords of the articles. We employed this formula: ((autocra* OR authorit* OR paternal* OR direct*) AND leader* AND perform*) We used autocra* to include words such as autocracy, autocratic, and autocrat; authorit* to include words such as authority, authoritative, authoritarian, and authoritarianism; paternal* to include words such as paternal and paternalistic; and direct* to include words such as directive and other possible variations. Moreover, we used leader* to include words such as leader, leaders, and leadership. Finally, we used perform* to include words such as perform, performance, and performances 2 Search 2.1

Search
We performed our search in SCOPUS and Web of Science. The total number of articles was 13,271. After excluding articles that were not written in English or not published in peer-reviewed academic journals, we obtained 9178 articles. After eliminating duplicates, we obtained a dataset of 6309 articles. Finally, after we limited the articles to those published in ABS journals ranked with 3, 4, or 4*, 765 articles remained in the dataset 1 3 Authoritarian leadership styles and performance: a systematic… Refine the sample Following Keupp et al. (2011) and Denyer and Neely (2004), we analyzed the title, abstract, and keywords of the extracted articles. We obtained a set of 73 articles. Then, we analyzed the contents of the articles, and we eliminated 19 articles. Therefore, the final sample included 54 articles 4 Analyze 4.1 Open coding Through open coding, we identified a set of concepts and categories from the articles' contents. This phase was characterized by high flexibility. We extracted concepts and categories that showed relevance for our research question Axial coding Through axial coding, we identified the relations between the categories. Following the paradigmatic model, the iterative process we performed during this phase let us understand the causal conditions among concepts, categories, phenomena, contexts, and consequences. We identified themes and subthemes 4.2 Selective coding Through selective coding, we integrated and improved the properties of the categories. Selective coding allowed us to achieve a higher level of abstractness. We identified the superordinate structures, i.e., generic and conceptual categories. These structures appeared to be the major themes of our content analysis 5 Present

Represent and structure the content
We structured the article's contents and descriptive statistics before structuring the entire article. This procedure allowed us to be more objective and stricter to the point

5.2
Structure the article We divided the article into five main sections. After the introduction, we described the methodology in detail. Then, we presented the relevant descriptive statistics that we considered valuable to answer the research questions proposed in the introduction. We presented the content analysis in the following two sections, performed through the grounded approach and the future research opportunities social responsibility. During the last decade, international business and area studies journals emerged as a new and relevant field interested in debating issues related to authoritarian leadership styles. The wide range of fields that debate the topic reflect its multifaceted nature. More specifically, the considerable number of viewpoints considered in discussing authoritarian leadership styles and their effects on performance reveal the complexity of the topic. Moreover, the influences of so many different fields on the scientific debate justify the vast number of contradictory results in empirical research. The considerable number of different scientific viewpoints demonstrates the phenomenological nature of the field. Therefore, the wide variety of areas and the increasing amount of contrasting empirical evidence should be seen as an opportunity and not as a limitation.

Evolution of authors' origins over time
The total number of authors included in the dataset is 148. The authors' origins evolved over time ( Fig. 4a and b). Between 1966 and 2010, most authors were from the USA (43, 66.15%), 6 were from Switzerland (although they were all concentrated in one article), five were from Canada (7.69%), four were from Israel (6.15%), and the rest were from Norway, Australia, France, or Finland. During the following decade, interest in the field shifted from Western to Eastern countries. Although most of the authors remained from the USA (29, 27.88%), China emerged with 23 authors (22.12%), Taiwan with 12 authors (11.54%), and Hong Kong with five authors. Together, Eastern countries contributed 40 authors, representing 38.46% of the total. Other countries represented during this period included the UK with ten authors (9.62%) and the Netherlands with four authors (3.85%).
To summarize, Western authors dominated the field of authoritarian leadership from 1966 to 2011, but Eastern countries emerged in the field during the last decade. The rapid development of Eastern economies, together with the increasing scientific competencies of the Eastern population, allowed these countries to participate in this debate. Comparing the current situation with the future conditions of the field in the next ten years would be interesting to verify whether this shift in scientific interest will be amplified by the continuous improvement of the Eastern scientific community's participation. Moreover, authoritarian leadership styles are typical in most Eastern businesses. Therefore, it would be interesting to verify whether research on authoritarian leadership by Eastern authors will improve the empirical knowledge of the field.

Evolution of statistical units' origins over time
A statistical unit is the unit of research observation: it is an entity-or set of entities-for which data are collected. The evolution of statistical units' origins over time was similar to the evolution of the authors' origins ( Fig. 5a and b). Many of the 47 empirical papers in our dataset involved collaboration between authors from 1 3 Authoritarian leadership styles and performance: a systematic… different countries (e.g., Chiang et al. 2020). Similarly, some papers used a sample of statistical units from different countries (e.g., Sutcliffe 1999; Wang et al. 2018).
Before 2011, more than 70% of the empirical papers employed statistical units from North America: 62.50% from the USA and 8.33% from Canada. Only three papers employed statistical units from Israel. After 2010, the data shifted. Only four studies employed statistical units from North America, and all of them were from b Fields (2011Fields ( -2021 1 3

Content analysis
In this section, we present the content analysis of the articles included in our dataset. We start with a conceptual map based on the work of Menz (2011) (Fig. 6). After that, we show the most commonly used theories in the articles. Further, we identify the results of our grounded analysis of themes that emerged during the content analysis. Finally, authors' suggestions for future research are discussed. To design the conceptual map in Fig. 6, we identified the main aim of each article in the database. We thus revealed three of the most significant general components of the scientific debate, i.e., context, actions, and consequences/outcomes, and we grouped the evidence that emerged from the content analysis into these three elements. After that, we found that mediators and moderators of leadership effects on outcomes/consequences were analyzed in detail by the considered literature. Table 2 shows the employed theories that appeared more than once in the dataset. Moreover, we show the leadership styles considered by the articles in which we found the theories.

Theories employed in the extracted articles
Paternalistic leadership is associated with Confucianism. Chen et al. (2014 referred to two pillars of Confucianism, i.e., hierarchy and relationalism. The former pillar involves respect for superiors, and the latter indicates that individuals with close relationships tend to "exchange favors beyond instrumental purposes" (Chen et al. 2014: 799). Wei et al. (2016) considered Confucian concepts of reflection (si), heart (qing), and mind-heart (xin) to achieve a more comprehensive framework of leaders' compassionate actions. Specifically, through these three concepts, the authors integrated the framework of compassion with the Confucian perspective.
Attribution theory aims to explain how subordinates or followers react to their perception of their leaders' behavior. Kipnis et al. (1981) used attribution theory to study employee evaluation. In particular, they referred to the power usage model to highlight that managers' perceptions of "who is in charge of employee's behavior" (Kipnis, 1981: 324) are the most critical variable for managers' evaluation of their employees. Moreover, Schuh et al. (2012) used attribution theory to study how followers seek leaders' signals to perceive their behaviors. Finally, Wang et al. (2013) used this theory to understand the impact of leaders' gender on subordinates' performance. Chang et al. (2003) used path-goal theory to identify four categories of leadership: participative, supportive, directive, and achievement-oriented. Kahai et al. (2004) added path-goal theory to adaptive structuration theory to predict the effects of two kinds of leadership, i.e., participative and directive.
Finally, contingency theory and cognitive resource theory were among the most used theories. For example, Yun et al. (2005) used contingency theory to show the different effects of leadership styles that adapt to different trauma conditions. Murphy et al. (1992) used cognitive resource theory to relate leaders' technical training to group members' performance.

Identified themes
This paragraph describes the results of our grounded analysis of the 43 articles' contents. In particular, we show the effects of authoritarian leadership style on the different typologies of performance highlighted by the authors. Figure 7 summarizes the results, highlighting the differences within the four leadership styles analyzed in terms of authority, power, attitude towards followers, and effectiveness. In the following section, we start from the authoritarian leadership style, and finish with paternalistic leadership.

Authoritarian leadership style and performance
Authoritarian leadership exercises control and authority over followers, limiting their autonomy and self-determination, and is often associated with the "dark side" of leadership (Chiang et al. 2020). The articles considered in this literature review which were devoted explicitly to this leadership style are relatively recent, and, in most cases, authors present studies in which the analyzed samples comprise Asian participants.
Authoritarian leadership and team performance In general, results in regard to the effects of authoritarian leadership on performance are coherent. In particular, most  units' origins (2011-2021) of the publications in our sample showed that authoritarian leadership had adverse (e.g., Scully et al. 1994), or, in some cases, insignificant (e.g., Chen et al. 2017) effects. A positive relation between authoritarian leadership and team performance was found by Fodor (1976), but this is a relatively isolated result. Moreover, authoritarian leadership affects followers' organizational and relational identity, often reducing their intention to stay (e.g., Schaubroeck et al. 2017;Shen et al. 2019) and suppressing emotions (e.g., Chiang et al. 2020). In particular, authoritarian leaders' lack of any emotional side leads to negative team performance, as verified by Chiang et al. (2020), who recently analyzed this phenomenon in relation to three large public Japanese companies.
Authoritarian leadership and in-role and extra-role performance Two articles debated the relations between authoritarian leadership style and in-role and extrarole performance. Schuh et al. (2012) considered a sample of 114 Chinese subordinate-supervisor dyads, finding that the leaders' authoritarian behavior had adverse effects on the two typologies of performance in the context of highly transformational leadership. In the context of low transformational leadership, however, the effects were not significant.  2021) considered Turkish, Taiwanese, and US employees and found that, although in the Asian context people are more willing to tolerate authoritarian leadership styles, leaders' authoritarian behavior can compromise the quality of communication, which is a critical moderator of followers' performance. Schaubroeck et al. (2017) considered the Chinese high-tech field and its employees and supervisors, demonstrating that an authoritarian leadership style promoted the worst job performance within contexts in which subordinates' power distance was not accepted by subordinates. In contexts in which the power distance was accepted, authoritarian leadership had no significant effects on job performance. Moreover, Shen et al. (2019) monitored how the indirect adverse effects of authoritarian leadership style on job performance varied according to the traditionality of groups and found significant effects, at least for groups with a low level of traditionality.

Authoritarian leadership and job performance
Authoritarian leadership style limits followers' self-efficacy and proactivity, as demonstrated by Li et al. (2019) in the context of the dyadic relations between Chinese leaders and followers. Moreover, an authoritarian leadership style can limit innovation and creativity (Lee et al. 2019).

Autocratic leadership style and performance
Although autocratic leadership shares its main characteristics with authoritarian leadership, it is less destructive and it is task-oriented; therefore, it tends to be associated with a higher level of acceptance by followers. In the set of articles we studied, autocratic leadership was not considered in Eastern Countries, although some  Authoritarian, autocratic, directive, and paternalistic leadership styles authors hypothesized that it could be more accepted in such contexts (e.g., Misumi & Peterson 1985). In general, autocratic leadership style can have both positive and negative effects on performance, depending on the contextual conditions (De Hoogh et al. 2015). In 1971, Rosenbaum and Rosenbaum considered the effects of autocratic leadership on task performance. Through an experiment which involved US students, they verified that the effects of this leadership style were positive on performance in a highly stressed condition but not on satisfaction. Ten years later, Kipnis et al. (1981) analyzed the effects of autocratic behaviors on leaders' evaluation of their followers. Given this leadership style does not leave decision-making power to followers, it leads to a worse performance, both at team and individual level. This condition leads leaders to evaluate followers more poorly.
Finally, autocratic leadership, like authoritarian leadership, promotes better performance in contexts in which hierarchy is well-accepted by followers (De Hoogh et al. 2015). Nevertheless, autocratic leadership can compromise the team's psychological self-confidence, worsening performance.

Directive leadership style and performance
Directive leaders exercise discipline and control over followers but, instead of focusing on making demands, they focus on providing guidance and feedback (Chiang et al. 2020). A coherent baseline can be identified during the entire period of directive leadership development in the literature, from Shaw and Blum (1966) to Yi et al. (2021). In particular, the latter made a critical contribution to the interpretation of directive leadership and its effects on performance, identifying an inverse U relation between the two variables. Although Yi et al.'s (2021) article is not a review, its results represent one of the more effective summaries to have emerged in the worldwide empirical literature to date. In fact, most of the literature has concentrated on finding specific situations in which a directive leadership style can have negative and positive effects on different typologies of performance.
Although the positive effects of directive leadership on performance and satisfaction could be considered counterintuitive at first sight, a convincing reason for their existence was provided by Kahai et al. (2004) through an analysis of cognitive evaluation theory. In particular, "structures external to an individual, such as leader directiveness, have control and informational components" (Kahai et al. 2004: 94). The two components affect, respectively, the conformity of behaviors and perceived competence. The informational component allows a reduction of perceived role ambiguity and an increase in followers' trust in leaders and motivation (Kahai et al., 2004;Chen et al., 2017).
Few articles explicitly found situations in which directive leadership leads only to adverse performance effects. For example, in the context of car dealers in North Europe, Chang et al. (2003) identified that customer focus could be compromised by a directive leadership style, adversely affecting quantitative parameters of performance measurement (e.g., productivity and profitability).
Directive leadership and task performance When the directive leadership literature first began to develop, academic analysis was concentrated in North America, especially the United States. In particular, tests in the first years were performed through laboratory experiments that involved undergraduates, and directive leadership was already seen to produce positive and negative effects on performance depending on the context. Shaw and Blum (1966) analyzed the effectiveness of directive leadership for task performance by studying changes in the task structures, and their experiment found that directive leaders managed more structured tasks better.
Nevertheless, not all studies have supported this finding. For example, Kahai et al. (2004) found a reduction of the positive effects of directive leadership for highly structured tasks. Therefore, the level of task structure which causes directive leadership to be more or less effective on performance remains an open issue.
Directive leadership and team performance: highly technical contexts Another relevant example is the work published by Murphy et al. (1992) that analyzed highly technical contexts. In particular, they demonstrated how, in such contexts, directive leadership could be effective in terms of team performance, but it has to be associated with the technical knowledge of leaders. This finding was confirmed by Sutcliffe (1999), who analyzed performances in business process reengineering through quantitative methods and the participation of IT executives from North America and Europe. The highly technical content of the context determined the greater efficacy of directive leadership in assignment and task facilitation, objective definition, and team performance. These findings were confirmed more recently by Rahmani et al. (2018), who verified how projects characterized by high knowledge intensity need directive leadership to be completed effectively.
Moreover, Hansen and Nørup (2017) considered a Danish ICT context and found that the concurrent employment of directive and participative leadership was the best possible strategy for perceived performance during project implementation. Similar results have been presented in the past. For example, Tjosvold (1984) identified that directive leaders' warmth and sincere interest in followers' productivity were effective mediators of work-related performance.
Further information on highly technical circumstances was provided by Yun et al. (2005) in a US medical context. In particular, their work highlighted how the directive leadership style promoted different effects depending on specific contextual elements. The study showed that such effects changed depending on the seriousness of the problem and teams' experience, with positive effects found at high levels of problem seriousness and low levels of team experience.
The same context was considered by Tschan et al. (2006) in Switzerland, where the authors identified positive effects of directive leadership on team performance. The study was performed in two phases, and it was found, in particular, that directive leadership had an additive effect, showing more effectiveness in the second phase for groups that had already obtained good results in the first. Hmieleski and Ensley (2007), who analyzed the context of the fastest-growing US startups, verified that directive leadership is better adapted to dynamic and unstable 1 3

Directive leadership and team performance: unstable and dynamic contexts
Authoritarian leadership styles and performance: a systematic… contexts in which decisions should be taken immediately. This tendency was confirmed by Lorinkova et al. (2013), who involved some US students in their study and demonstrated that team performance in groups guided by directive leaders was positive in the short run and stabilized thereafter. At the beginning of project implementation, the context is unstable, tasks are not permanently assigned, and competencies are yet to be developed. In such contexts, directive leadership is more effective. Mossholder et al. (1990) had already found similar results by integrating time in their study of leadership. In fact, the longer the duration of dyadic relations, the less followers are affected by leaders' behaviors. In particular, directive leadership is initially more effective in complex contexts (e.g., engineering). Over time, followers become confident in their competencies, and non-directive styles are the best choice for satisfaction and job performance. Sanchez-Manzanares et al. (2020) further confirmed such effects in the Spanish context. Their study verified that when pressure is strong, directive leadership is more effective for team performance and adaptation.
Given that leadership is a highly phenomenological construct, not all the studies found significant relationships between directive leadership and team performance. For example, Faraj and Sambamurthy (2006), who also analyzed a technological context (i.e., information systems development projects) in the USA, found nothing to support this relationship.
Directive leadership and team performance: Asian contexts Although there has been a particularly notable recent contribution from Eastern authors to the literature on directive leadership, interesting works have been published by Asian academics previously. These publications suggested that this typology of leadership style is more accepted in Eastern countries than in Western areas. Nevertheless, we identified a tendency among Eastern authors to suggest that smoothing the directive side of leadership can foster better performance. For example, Sagie (1996) developed a study which involved the participation of students in Israel and underlined that directive leadership produced positive results but that, at the same time, when leaders used a more communicative approach to define goals in a participative way, the positive effects on team performance increased.
An interesting comparison can be made between Somech (2006) and Hmieleski and Ensley (2007), who considered the effects of directive leadership on team performance, starting from different levels of team homogeneity. In this sense, interpreting directive leadership as a mediator or moderator of such an effect, Somech (2006) considered the Israeli context and analyzed in-depth the effects of directive leadership style on in-role performance and team innovation, considering different levels of team homogeneity. In particular, a more homogeneous team, in terms of professional backgrounds and abilities, can put more pressure to achieve conformity. In this case, directive leadership promotes better performance. In contrast, an analysis of the US context by Hmieleski and Ensley (2007) highlighted that top management team heterogeneity could lead to positive results if a directive leadership style is adopted. Therefore, to achieve better performance, directive leadership should be associated with top management team heterogeneity and low-level team homogeneity.

Paternalistic leadership and performance
The paternalistic leadership style is traditionally described as a construct composed of three dimensions: authority, morality, and benevolence. Nevertheless, Wang et al. (2018) published an innovative interpretation of this leadership style. In particular, they stated that previous literature joined three separate styles: benevolencedominant paternalistic leadership, in which benevolence dominates over authority; authoritarianism-dominant paternalistic leadership, in which authority dominates over benevolence; and classical paternalistic leadership, in which no dimension dominates. Therefore, the definition and foundation of paternalistic leadership style can be considered an open issue in the literature.
In general, paternalistic leadership is based on Confucian philosophy, which is particularly studied in Eastern contexts, in which it is present and appreciated (Shen et al. 2019). In the set of articles we studied, almost all the articles analyzed Asian statistical units, and there was a general tendency to consider paternalistic leadership an effective style in terms of performance. Chan et al. (2012) and Chou et al. (2015) considered the effects of paternalistic leadership on task performance. The former concentrated on the authoritarian and benevolent components of this leadership style in relation to Chinese supervisor-subordinate dyads. The benevolent dimensions were shown to be a moderator of the adverse effects caused by the authoritarian dimension on task performance, organization-based self-esteem, and organizational citizenship behavior toward the organization. The latter, which focused on the Taiwanese military, considered the authoritarian and moral dimensions of paternalistic leadership and found that, in such a context, high levels of both dimensions were the most widespread and appreciated by followers.

Paternalistic leadership and task performance
Paternalistic leadership, job performance, and trust Numerous studies have considered the effects of the paternalistic leadership style on work and job performance, and Wu et al. (2012) studied such effects by analyzing supervisor-subordinates dyads in the Chinese context. Although paternalistic leadership has already been associated with positive performance, the researchers identified international justice as a possible mediator of this relation, both for job performance and organizational citizenship behavior. Wang et al. (2018), who compared dyads in Taiwanese and US contexts through their innovative vision of paternalistic leadership, verified that if the authoritarian dimension dominates leaders' style, the effects on job performances are adverse. In contrast, if benevolence is dominant, or if there is an equilibrium between the two dimensions, the effects are positive. The same results were obtained by Hiller et al. (2019), who explained that adverse effects could extend to organizational citizenship behavior, creativity, turnover intention, and counterproductive attitudes. Chen et al. (2014) found significant influences of all three dimensions of paternalistic leadership for in-role and extra-role performance. Their results were similar to those of previous studies: the authoritarian dimension negatively affected the two typologies of performance differently from the other two dimensions. Moreover, affective trust was a significant mediator between benevolent and moral dimensions and role performance.
The meta-analysis carried out by Legood et al. (2020) demonstrated that affective trust could effectively predict followers' performance. Moreover, this study remarked that specific conditions exist which foster the effectiveness of the paternalistic leadership style. In particular, trust is promoted by a paternalistic leadership style when the power distance is low. When the power distance increases, however, the benevolent dimension should be more and more dominant to ensure a positive work-related performance.

Directive leadership vs. empowering and participative leadership styles
Numerous articles have debated the differences between directive and empowering leadership styles. Specifically, empowering leadership is a style focused on "sharing power with employees and increasing their responsibility and autonomy to perform their work" (Sanchez-Manzanares et al. 2020: 840). Most articles debating this comparison state that directive and empowering leadership styles have different and significant effects on performance, with limited exceptions (e.g., Faraj & Sambamurthy 2006).
In stressful contexts, directive leadership showed higher effectiveness in regard to team adaptation and performance (Sanchez-Manzanares et al., 2020). In contrast, in less complicated situations, empowering leadership has a positive effect on team performance (Yun et al. 2005).
Empowering and directive leadership were also compared according to levels of team homogeneity. In particular, Hmieleski and Ensley (2007) examined the US startup context. They found that, in unstable environments, a directive leadership style was more effective for firm performance when top management teams were heterogeneous, while empowering leadership positively affected performance when top management teams were homogeneous. In stable environments, however, the effects were the opposite.
Other studies performed in similar contexts gave different results in regard to comparisons of directive and participative leadership styles. Rahmani et al. (2018) found that directive leadership is more effective than participative leadership in fields with high knowledge intensity, whereas Hansen and Nørup (2017) found that employing a directive and participative leadership style combination is the best strategy for implementing projects in ICT environments. The same results were obtained in an experiment performed by Sagie (1996).
Moreover, through experiments, Sauer (2011) found that the effectiveness of leaders depends on their position in the firm. In particular, low-status leaders are perceived as more effective if they adopt a directive style, while high-status leaders are perceived as more effective if they adopt a participative leadership style. Somech (2006) found that when groups are functionally heterogeneous, the participative leadership style is more effective than the directive style, which is ineffective in functionally homogeneous groups. Directive leadership can reduce role ambiguity and, therefore, increase the quantity of communication. Thus, directive and participative leadership can positively affect group performance and satisfaction (Kahai et al. 2004).

Further research opportunities
This section presents further research opportunities that selected articles' authors identified in their research. We found four main areas of interest (Table 3). First, replications of studies in different cultural contexts were widely suggested by authors. Second, further research opportunities examining novel and multiple leadership styles were identified. Third, emotions and perception were suggested as promising further research opportunities. Finally, authors identified the need to analyze the power of moderator and mediator variables influencing the relationship between leadership styles and performance.

Discussion
In this paper, we performed an SLR to achieve a better understanding of authoritarian leadership styles on the different typologies on performance considered in the literature. Although the literature presents considerable evidence highlighting the need for leaders to adapt to different contexts (e.g., Yun et al., 2005;Chiang et al., 2020), our findings showed that, when leaders limit the authoritarian component of their styles, there are more contexts in which the considered leadership styles can be effective. The level of authority, discipline, and control over followers decreases as leaders move from authoritarian to paternalistic leadership styles, passing through autocratic and directive styles.
There is a general consensus among researchers that the authoritarian style is the worst leadership style for performance. Differenly, the autocratic leadership style is more task-oriented and tends to be more accepted by followers (Chiang et al. 2020). It has positive effects on performance in very stressful environments. In contrast, the directive leadership style has been shown to produce positive effects on performance in a considerable number of contexts, and it has been most frequently demonstrated to be an effective strategy for numerous typologies of performance in unstable, dynamic, and highly technical environments. Paternalistic leadership has been shown to be the most balanced style. Its three dimensions (i.e., authority, morality, and benevolence) lead to better effects on performance, especially when the benevolent and moral dimensions dominate the authority one.
Finally, although authoritarian leadership styles are more accepted in Eastern countries (e.g., Karakitapoğlu-Aygün et al. 2021), the literature showed a tendency to ask that the "dark side" of leadership style be smoothed, even in Asian contexts. Smoothing the authoritarian side of leadership could be obtained through better communication (e.g., Karakitapoğlu-Aygün et al. 2021;Sagie 1996) or through leaders' sincere interest in their followers' lives and productivity (Tjosvold 1984).

3
Authoritarian leadership styles and performance: a systematic… In the following section, we present the limitations of our work, propose a research agenda for further in-depth analysis in line with our study, and discuss the practical implications of our paper.

Limitations and research agenda
The principal outcome of our study is the finding that a considerable number of mediators, moderators, and specific conditions can change the effects of authoritarian leadership styles on performance. In particular, the leadership style appears to be a complex condition that cannot be superficially ascribed to a single and peculiar style detached from other leadership models that supervisors may apply. It is, instead, a changing condition: leaders are often able to change their degree of authority in line with objectives, followers' behaviors, and other specific conditions (Kahai et al. 2004). For example, paternalistic leaders can increase or decrease leadership dimensions (i.e., authoritarian, benevolence, and morality) depending on the context, actions, subordinates' behaviors, and desired outcomes.
Given that a discussion about single and detached leadership styles does not make sense in light of the dynamism of the new globalized and hyperconnected markets, the study of hybrid leadership styles is still open and promising. The indepth analysis of specific circumstances in which leaders can change the degree of a specific dimension of their leadership styles is a critical stimulus for future researchers. In particular, authoritarian leaders are suitable units of analysis: the literature widely recognizes that more democratic leadership styles (e.g., empowering) positively affect performance. Instead, authoritarian leaders have to increasingly adapt their style, which collides with the changing conditions of jobs, markets, and diffuse beliefs about the negative impact of exaggerated authority (Sauer 2011). Therefore, in future research, examining authoritarian leadership styles will probably be the key to achieving an in-depth understanding of hybrid leadership styles.
Moreover, we encourage future researchers to overcome two limitations of our paper. The first relates to the conceptual map we drawn in Fig. 6. The thematic map shows three main groups of evidence: context, actions, and consequences/outcomes. These three dimensions are connected in a linear relationship that starts from the context and produces the effects on performance. Nevertheless, Dourish (2004) suggests that in fields where the phenomenological nature of results is undeniable (as it is for leadership), context is much more than a starting point. In particular, "from a phenomenological perspective […] context does not describe a setting; it is something that people do. It is an achievement, rather than an observation; an outcome, rather than a premise" (Dourish 2004: 22). Therefore, we encourage further research to find theoretical and empirical evidence of performance effects on leadership. In particular, are leaders willing and able to change their style depending on previous performance?
The second limitation of our work is that the comparison we made between authoritarian and non-authoritarian leadership styles ( §5.2.5) was not the final aim of our article. It was a secondary result. However, it underlined a critical starting point for further research opportunities. In particular, we encourage What are the effects of power distance in leader-follower agreement in contexts in which specific cultural dimensions vary (e.g., assertiveness)?
Karakitapoğlu-Aygün et al. (2021) What are the effects of abusive supervision outside China? Li et al. (2019) What is the role of relational identification and workgroup cultural values in linking authoritarian leadership to employee performance through dissimilar cultures? Shen et al. (2019) What is the role of paternalistic leadership in the U.S. or other Western settings? Wang et al. (2018) What are the effects of further tests of adaptability of classical paternalistic leadership in Western contexts? Wang et al. (2018) What are the effects of directive achieving on performance in non-Chinese contexts that share similar characteristics (paternalistic leadership)?
Chen et al. (2017) Are there additional sites on compassion in different industries or Countries? Wei et al. (2016) What are the effects of autocratic leadership on performance in highpower-distance cultures?
de Hoogh et al. (2015) What are the effects of paternalistic leadership profiles in non-Taiwanese contexts and across cultures? Chou et al. (2015) What are the effects of affective trust on performance in non-Chinese contexts that share similar characteristics (paternalistic leadership)? Chen et al. (2014) What are the effects of authoritarianism, benevolence, and subordinates' organization-based self-esteem on performance in non-Chinese cultures and work settings? Chan et al. (2012) How do cultural values relate to differences in leadership perceptions?

Call for research on novel and multiple leadership styles
What are the relative influences of established styles in novel leadership models? Legood et al. (2020) What are the effects of leaders' ability to engage in different leadership styles in the context of adaptive teams?
Sanchez-Manzanares et al. (2020) What are the effects of directive-achieving leadership styles on performance? Chen et al. (2017) What are the effects of leaders' ability to alternate between higher and lower levels of autocratic leadership?
de Hoogh et al. (2015) What are the effects of multiple types of leaders on long-term teams? Lorinkova et al. (2013) What are the effects of employing experimental designs to manipulate abusive supervision and authoritarian leadership? Li et al. (2019) How do the effects of directive-achieving leadership evolve? Chen et al. (2017) What are the effects of transformational leadership with different leadership behaviors?

Emotions, perceptions and personality traits
What are the effects of emotion suppression and exhaustion in different experimental conditions? Chiang et al. (2020) 1 3 Authoritarian leadership styles and performance: a systematic… future researchers to explore in much more detail the differences among leadership styles in their effects on performance. We call for research reviewing a broader range of leadership styles to highlight the main differences in their outcomes. We are aware that this is a demanding objective. Therefore, in our opinion, meta-analyses should help organize the considerable number of papers published on the topic. In particular, a restriction in the time range of articles extracted could support researchers in limiting bias and reducing the time needed for the analysis. What are the differences between the suppression of positive and negative emotions advocated by authoritarian leadership? Chiang et al. (2020) What are the effects of trust on senior leadership? Legood et al. (2020) What is the relationship between right-wing authoritarianism and a desire for submissive status? Hing et al. (2007) What are the effects of pairing leaders who are driven by dominance over others and lack empathy with followers who are driven by obedience to authority and lack independence? Hing et al. (2007) Can electronic brainstorming be responsible for the negative relationship between participation and satisfaction? Kahai et al. (2004) What are the effects of participation on performance via satisfaction? Kahai et al. (2004) What variables can tap the processes that translate the influence of perceived powerlessness on proactivity?
Li et al. (2019) Can pseudo-transformational leadership mediate the dark side of leaders' personalities and undesirable employee outcomes? Schuh et al. (2012) What are the effects of participants' emotion suppression and exhaustion in different experimental conditions?

Moderators and mediators
Does group traditionality moderate performance, authoritarian leadership, and relational identification?
Shen et al. (2019) Do team personality composition, group cohesion, team empowerment, and shared leadership moderate abusive supervision effects? Li et al. (2018) What are the specific power struggles that moderate autocratic leadership effects?
de Hoogh et al. (2015) What are the potential moderators in the financial performanceleader behavior relationship? Scully et al. (1994) What mediators explain classical paternalistic leadership? Wang et al. (2018) Do power struggles mediate the relation between leader consideration and team psychological safety?
de Hoogh et al. (2015) Does trust mediate the relationship between paternalistic leadership and organizational citizenship behavior? Legood et al. (2020) Do leader-member exchange theory and trust mediate psychological or relational processes?

Managerial implications
While the theoretical implications of our study are clear, since we fill the gap concerning the effects of authoritarian leadership styles on performance, this review also has relevant practical implications. In particular, we identified three groups that could benefit form the managerial implications: leaders, top management teams, and recruiters and human resources management (HRM). Leaders should be aware of their leadership styles (Chiang et al. 2020;Karakitapoğlu-Aygün et al. 2021). Given that under specific conditions, authoritarian leadership is effective, leaders have to be trained to understand the nature of their behaviors. To achieve this demanding objective, they have to gain indepth knowledge of each leadership style's positive and negative effects. As a consequence, they should be able to limit the degree of their authority and their centralizing tendency. There are several opportunities to help leaders achieve this goal. For example, HRM can design jobs with autonomous features (Li et al. 2019) and encourage regular feedback between superiors and subordinates (Karakitapoğlu-Aygün et al. 2021).
Top management teams should design the work environment to enhance creativity, share ideas, incentivize collaboration and information sharing, and encourage participation (Lee et al. 2019;Shen et al. 2019) to stimulate authoritarian leaders to reduce the power distance between them and subordinates. Planning work activities in line with democratic ideas can help followers identify with their leaders. Moreover, supervisors should be trained to clearly express their projects for their subordinates, especially in terms of their learning and professional achievements (Chen et al. 2017). Furthermore, leaders should be subjected to psychological training to improve their ability to understand the structure of their teams and the personality of their subordinates.
Finally, in the last decade, the literature has reconsidered the strategic role of HRM (e.g., Pizzolitto and Verna 2020;2022). Given the evidence that emerged during this review, the fundamental effects of strategic HRM interventions can dramatically affect leadership and, consequently, performance. Recruiters should select managers depending on the leadership styles required by the specific conditions of workgroups, markets, and business needs. For example, evidence in the literature suggests that changing and complex conditions can be better managed through directive leaders (e.g., Lorinkova et al. 2013;Sanchez-Manzanares et al. 2020). In calmer situations, more empowering leadership styles are more effective. Therefore, recruiters have to be careful in their selections, considering the training and compensation needed for adapting managers' leadership styles to business needs. In specific situations, hiring a directive leader can be an efficient choice (e.g., Yun et al. 2005;Lorinkova et al. 2013).
Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.