While in the employ of one of the leading conglomerates (International Business Machines [IBM]), Geert Hofstede conducted a survey among employees in more than 72 countries between 1968 and 1973 to understand the cultural orientation of people in different countries (Hofstede, Hofstede, & Minkov, 2010). His extensive and evolving research findings led to the development of a framework with five cultural dimensions. The original five dimensions of Hofstede’s framework (1980) are power distance, uncertainty avoidance, masculinity–femininity, individualism–collectivism and long-term orientation (LTO), which was added as the last dimension (Minkov & Hofstede, 2012). Power distance describes the extent to which inequalities are accepted among the people of a society, with people in countries high in power distance more likely to accept and expect differences in power among them, whereas in countries with low power distance, people expect that the differences in power should be minimised. Uncertainty avoidance reflects the extent to which people in a society try to avoid situations that do not provide certainty; the masculinity–femininity dimension reflects how societies distribute roles based on gender; the individualism–collectivism dimension describes the extent to which individuals in a society are either integrated into groups or encouraged to be independent; long-term orientation refers to the extent to which a community or society embrace long-term and futuristic commitments, as opposed to commitments that are short-term focused, thereby differentiating societies that are able to embrace change more rapidly than the others.

Although Hofstede’s framework has been challenged for various reasons, including the validity and representativeness of his data (Baskerville, 2003; Javidan, House, Dorfman, Hanges, & De Luque, 2006; Minkov, 2018), his work continues to be referenced to this day (Areiqata et al., 2021; Gallego-Álvarez & Pucheta-Martínez, 2021; Kaasa, 2021; Olatunji, Makhosazana, & Vezi-Magigaba, 2021; Olowookere, Agoha, Omonijo, Odukoya, & Elegbeleye, 2021). The various disciplines that reference his work include cross-cultural management (Beugelsdijk, Kostova, Kunst, Spadafora, & Van Essen, 2018; Klasing, 2013), country-level culture (Beugelsdijk, Kostova, & Roth, 2017) and international business (Kirkman, Lowe, & Gibson, 2006).

The GLOBE research focuses on the interplay between culture and leadership effectiveness, based on extensive work conducted in 951 organisations in 62 countries (Hoppe & Eckert, 2012; House et al., 2014). This research provides a view of (1) how cultures are different from or similar to one another in their perceptions and expectations of leadership effectiveness and (2) the definition or perception of what constitutes an outstanding leader. Six of the nine GLOBE cultural dimensions were built on the work of Hofstede House, Hanges, Javidan, Dorfman and Gupta (2004), while some of the scales the GLOBE used to measure cross-cultural differences (uncertainty avoidance, power distance, individualism) are designed to reflect the same constructs as Hofstede’s dimensions. The GLOBE initially presented nine dimensions, which were later refined and reduced to six, all of them describing how leadership is expected or experienced across different cultures (House et al., 2014). The six dimensions are charismatic or value-based, team-oriented, self-protective, participative, humane-oriented and autonomous (House et al., 2014).

According to the GLOBE study, cultural similarity is greatest among societies that constitute a cluster, while cultural difference increases the further the clusters are apart (Center for Creative Leadership, 2012). For example, the countries from the sub-Saharan region in Africa (Nigeria, Zambia, Namibia, Zimbabwe and South-Africa¹) constituted a cluster and as a result, more cultural similarities would be seen in these societies. This African cluster, for example, would practice leadership differently from countries in parts of the Global North, such as Eastern Europe (e.g. Russia, Poland, Georgia, Greece, Hungary, Albania, Slovenia). It was also recognised that different cultures conceptualise outstanding leaders differently, and this difference is because of cultural upbringing and expectations about what entails good leadership (Center for Creative Leadership, 2012). To that end, there is acknowledgement that the location and culture within which leadership occurs provides important context for consideration on leadership discourse (Eyong, 2017).

Hofstede and the GLOBE cultural research both proved to be similar in the conceptualisation of the cultural dimensions.

Since the launch of GLOBE, Western leadership paradigms that have long been embedded in leadership practice globally are being questioned (Bolden & Kirk, 2009; Eyong, 2017). It has been observed that the global leadership practices of the 18th and 20th centuries were influenced by Eurocentric theories and narratives that do not reflect the indigenous dimensions and contexts that are critically important to understanding leadership (Eyong, 2017; Wanasika, Howell, Littrell, & Dorfman, 2011) outside Eurocentric perspectives. As a result, there has been a growing movement suggesting that leadership should be understood within the context in which it is practised (Avolio, Walumbwa, & Weber, 2009; Humphreys, Zhao, Ingram, Gladstone, & Basham, 2010; Jones, Lefoe, Harvey, & Ryland, 2012; Muchiri, 2011; Posner, 2012). This emerging research, along with the current interest in understanding the role of culture and context in leader effectiveness, presents diverse findings, thus further widening the gap on the definition of leadership – especially leadership in the African context (Lerutla & Steyn, 2017).

Africa, as a region, has drawn interest in leadership research, given the dynamic nature and uniqueness of the continent (Lerutla & Steyn, 2017). Researchers (e.g. Bolden & Kirk, 2009; Chasi & Levy, 2016; Chatbury, Beaty, & Kriek, 2011; Littrell & Nkomo, 2005) argued that Africa has its own values and practices and that these should be integral to the definition of leadership and its implementation on the continent. For example, Bolden and Kirk (2009) argued that Africans aspire to leadership based on humanity, a leadership that is inclusive and that values individual differences, along with a desire to serve others. They argue that there is a pattern unique to Africa, which can be associated with the belief system and philosophical principles of ubuntu – a term that represents African humanism and that is often translated as ‘I am because we are’. Ubuntu is a fundamental philosophy that governs existence and social relations in sub-Saharan Africa (Mangaliso, 2001; Mbigi, 2000; Wanasika et al., 2011). It is premised on respect for the dignity of people, reciprocity in social relations and a desire for tolerance and forgiveness.

In other research, Gumede (2017) took the position that African leadership is grounded on Afrocentric histories, philosophies, epistemologies and other pan-Africanist ideologies within the evolving African cultures. Metz (2018:p. 42) also talked about an Afro-communal ethic type of leadership, unique to Africa, based on the belief that ‘one should relate communally and enable others to commune’. Afro-communal leadership is associated with servant leadership, although not in a way understood in the Western context; servant leadership is distinguished as leaders’ interest in meeting the needs of others, with emphasis placed on communal relationships (Metz, 2018). In his address to a group of leaders, Khoza (2012) linked leadership in Africa to humanness and further argued that the human style of leadership generated by African humanism puts people first.

In his research, Mbigi (2000) has pinpointed five core values associated with African leaders: respect for the dignity of others, group solidarity, teamwork, service to others and the spirit of harmony and interdependence. These values have been validated in other research findings. For example, in a study conducted by Flotman and Grobler (2020), it was established that servant leadership has an impact on organisational behaviour within the South African and African context. Muller, Smith and Lillah’s (2019) study found that the ubuntu values of solidarity and servant leadership had a significant and positive influence on employee engagement and organisational performance. Molose, Goldman and Thomas’s (2018) research outcomes indicated that the ubuntu collective values of compassion, group solidarity, respect and dignity for all are beneficial to workplace commitment and performance. In their study, they recommend that ubuntu should be incorporated in management approaches to facilitate workforce commitment and team performance. The GLOBE findings also reported that sub-Saharan respondents mostly supported aspects such as human-oriented, charismatic and/or value-based and team-oriented leadership (Hoppe & Eckert, 2012).

The FRLT originated from the work of James Burns in 1978, which posited that leaders are either transformational or transactional, thereby positioning the two styles as being at opposite ends of a continuum (Burns, 1978). Following extensive research by Bass (1990), this original work was expanded to argue that the best leaders are both transformational and transactional. With further research, Avolio and Bass (2005) posited that to achieve organisation success, a paradigm shift was required to understand the overall role that leaders play in influencing followers to go beyond self-interest for the greater good of the entities with which they are associated (Khan, Nawaz, & Khan, 2016). The crux of their argument was that leadership is neither transactional nor transformational but rather that it happens on a continuum. The leadership constructs that form the core of the FRLT are the transactional, transformational and laissez-fair styles (Avolio, Bass, & Jung, 1999; Bass & Avolio, 1994), which are briefly outlined in the following paragraphs.

The population comprised all leaders at all organisations operating in South Africa. A convenience sample was drawn from the subordinates of these leaders. The respondents were recruited by students pursuing a Master of Business Leadership (MBL) degree at the Graduate School of Business Leadership at the University of South Africa (GSBL). The students, through their respective employer organisations, gained access to the respondents. The students then acted as fellow researchers, collecting data from 1140 respondents across 19 different organisations. The respondents were selected on a random basis from the personnel records in the participating companies.

Demographic information was collected, consisting of the gender, race and age of the respondents, as well as of the leaders. Those identified as black were deemed to be of African heritage and those who identified as white were deemed to be of Western descent. Responses from those of Asian descent and those identified as mixed race were deleted for the comparative analyses.

The FRLT elements were assessed with the Pearce and Sims (2002) instrument, with 20 TFL items, 11 TSL items and five LFL items. Responses to the items were recorded on a five-point scale, with ‘Definitely not true’, ‘Not true’, ‘Neither true nor untrue’, ‘True’ and ‘Definitely true’ as options. A sample item from the TFL section is ‘My leader provides a clear vision of who and what our company is’; from the TSL section, ‘My leader will recommend that I am compensated well if I perform well’; and from the LFL section, ‘My leader allows performance to fall below minimum standards before trying to make improvements’.

Leadership effectiveness was measured, using an instrument developed by Cicero, Pierro and Van Knippenberg (2010) and comprised four items. Items responded to a seven-point scale, ranging between ‘Strongly disagree’ and ‘Strongly agree’. A sample item from this instrument reads as follows: ‘My leader influences my level of commitment effectively’. The instrument had acceptable levels of reliability, with a Cronbach’s alpha equalling 0.83 (Cicero et al., 2010). Low scores on leadership demonstrate a low level of that particular behaviour style, for both leader effectiveness and leadership style.

IBM’s Statistical Package for the Social Sciences (SPSS) was used to perform the data analysis. Demographic data were calculated, mainly focusing on identifying the cultural heritage of the respondents, as well as that of the managers they were evaluating. White respondents and managers were deemed as representative of Western heritage and black respondents and managers as representative of African heritage.

Reliability data were collected in the form of Cronbach’s alpha. A value larger than 0.7 was deemed as sufficient proof of reliability (Nunnally, 1978). Cronbach’s alphas were calculated for each of the subscales of the FRLT, as well as for the leadership effectiveness instrument.

Validity data were generated through exploratory factor analyses, and the results revealed a simple structure in the data (Tabachnick & Fidell, 2019), with few cross-loadings and items loading in accordance with the structure of the instrument were deemed as evidence of validity. This analysis was performed only on the FRLT items, as the leadership effectiveness instrument consisted of only four items and was deemed unidimensional.

Means were calculated and presented and mean differences were then calculated using ANOVA. The cut-off score for statistical differences between means was set at 0.05. If mean differences were found, Cohen d-values were calculated. Cohen (1988) indicated that d-values up to 0.2 represent a negligible effect, between 0.2 and 0.5 a small effect, between 0.5 and 0.8 a medium effect and beyond 0.8 a large effect.

Correlations between the different leadership styles and leadership effectiveness were calculated per race. The interest in this study was not the size of the correlation but rather the differences of these correlations across race. Z-observed scores were calculated to determine if these correlations differed significantly from each other (Field, 2013). Z-observed scores higher than (+/–) 1.96 were interpreted as indicative of a significant difference between the correlations, at p < 0.01 (Pallant, 2020). When z-observed scores were smaller than (+/–) 1.64, it was assumed that the differences in the correlations were not significant.

Permission to conduct the research was granted by the Research Ethics Review Committee of the Graduate School of Business Leadership under approval number 2016_SBL_003_CA. Students who assisted with the data collection were informed of the conditions under which the data would be used. They were required to obtain formal written approval from relevant executives in the organisations where they collected the data. All potential respondents received a comprehensive information sheet explaining the nature of the research and the way the data would be used. Participation was both voluntary and anonymous, and it was explained to the participants that completing and submitting the questionnaire would be deemed as consent to participation.

Responses from 1140 employees were captured. Of these, 573 (50.3%) were men and 567 (49.7%) were women. Black people were the dominant group, with 762 respondents (66.8%), followed by white people with 206 respondents (18.1%) and smaller numbers of mixed race (116 or 10.2%) and people of Asian descent (56 or 4.9%). The age of the respondents varied between 20 and 64 years, with a mean of 38.62 and a standard deviation of 9.364.

The respondents were asked questions about their leaders. In total, 702 (61.6%) of the leaders were identified as men and 437 (38.3%) as women. The race composition of the leaders reported on were as follows: 655 (57.5%) black, 350 (30.7%) white, 90 (7.9%) mixed race and 45 (3.9%) of Asian descent. The age of the leaders reported on by the respondents varied between 23 and 70 years, with a mean of 44.23 and a standard deviation of 8.380.

The TFL scale showed high levels of reliability, with a Cronbach’s alpha of 0.942 (20 items). The Cronbach’s alpha for TSL was 0.750 (11 items), 0.834 (5 items) for LFL and 0.951 (4 items) for the measure of leadership effectiveness.

The factorial validity of the leadership style instrument was tested. An acceptable Kaiser–Meyer–Olkin measure of sampling adequacy score of 0.939 and a Bartlett’s test of sphericity approximation with a significant (p < 0.001) chi-square (df = 630) of 27070.982 indicated that the data on the 26 items were suitable for factor analysis. A five-factor solution provided the best fit and declared 60.5% of the variance in the data set. The items from TFL loaded on factors 1 and 3, while the items from TSL loaded on factors 2 and 4. No cross-loadings larger than 0.4 occurred across these two groups of factors. All the items from LFL loaded on factor 5, with one item having a significant cross-loading on factor 4 (which forms part of TSL). The observed fit of the data suggested that, at a factorial level, the measurement of leadership was valid.

The mean scores for leadership style for black and white leaders, as assessed by their subordinates, are presented here. Low scores on leadership show low levels of that behaviour style and, similarly, low scores on effectiveness reflect low levels of leader effectiveness.

From Table 1, it can be observed that, apart from LFL (F(1,1001) = 5.352; p = 0.021), the mean leadership scores did not differ from each other statistically. In Table 2, it can be read that the mean score on LFL for black people was 2.517, while for white people it was 2.378. The Cohen’s d-value ((2.517–2.378)/0.906) was 0.058, which suggests that, at a practical level, the difference was negligible.

From Table 3, it can be observed that the means did not differ significantly. This is also evident from Table 4, where the mean score for black people was 5.084, and that of white leaders was 5.037.

Table 5 reports the correlation between the different leadership styles and leadership effectiveness and the data regarding the differences between these correlations, which is reflected by z-observed values.

In Table 5 numerically, black leaders who display TFL are perceived to be more effective than white leaders and white leaders who display TSL are judged as more effective than black leaders. There is a negative correlation between LFL and leadership effectiveness, and this correlation, in numerical values, is stronger for black than for white leaders. However, when testing for the significance of the differences between the correlations, the z-observed values indicate that these differences are not statistically significant.