Creating job satisfaction in Africa: A case study from Mozambique

Abstract

How to set up successful multicultural human resource management programmes remains a contentious issue. This study examined variables that influenced job satisfaction in the workforce of a multinational company operating in Mozambique. It took a mixed methods approach and analysed case study data using a combination of content, thematic, descriptive and statistical techniques. The results indicate that the level of job satisfaction of these employees was comparable to that of employees in former Soviet Bloc countries; however, it varied significantly across levels of employee. The results also demonstrate that, to be more successful, ‘foreign’ management practices should be adapted to the local culture and, further, that humanism in the African workplace can positively mediate a traditional instrumentalist rewards system.

Keywords:

• job satisfaction
• human resource management
• Mozambique

1. Introduction

Multinational companies investing in Africa often inappropriately import and implement foreign based instrumentalist type human resource management (HRM) programmes that reflect a contract based view of an employee as merely a resource (Jackson, 2002). The management practices of these companies are often unquestioningly accepted by local economies that may be desperate to secure foreign investment. Needless to say, whether local employees are open or resistant to new business partnerships, they soon start to question HRM practices that are alien to their culture, while the parent company cannot understand why the indigenous workforce are not more grateful for their newfound prosperity (Gomez-Mejia & Palich, 1997; Sousa-Poza & Sousa-Poza, 2000; Hofstede, 2001; Jackson, 2002; Pitcher, 2006; Carmichael & Rijamampianina, 2007; Horwitz, 2007). Rijamampianina & Carmichael have indicated that organisations fail more often because of insufficient consideration of cultural and, by implication, human factors than because of strategic breakdown: ‘To ignore the human factor with all its associated elements is, in the longer term, sure to invite failure’ (2005:93).

The recent experiences of multinationals operating in Africa underline the need for a better understanding of local cultures and values, particularly when it comes to developing HRM policies (Pitcher, 2006; Carmichael & Rijamampianina, 2007; Dimba & K'obonyo, 2007). It has been observed that cultural values in particular need to be respected if a workforce is to be content and productive (Ogbonna & Harris, 2002). This HRM problem is underlined by the growing corporate militancy of workers (Pitcher, 2006; Luiz, 2007), as well as a drop in productivity, poor motivation, increased levels of alienation, labour unrest and strikes (Human, 1996; Pitcher, 2006; Carmichael & Rijamampianina, 2007; Dimba & K'obonyo, 2007; Horwitz, 2007).

In Mozambique, a central strategy of the Frelimo government after 1992 was to promote foreign investment. One such investment was the Mozal project, an aluminium smelting operation situated in Maputo, the capital city of Mozambique. The Mozal Company, the focus of this paper, is owned by BHP Billiton (47%), the Mitsubishi Corporation (25%), the Industrial Development Corporation of South Africa (24%) and the Mozambique Government (4%). The company initiated operations in June 2000 and by 2001 was employing 750 personnel. The newly employed workforce was young and had no experience of working in a high technology production environment. From the outset, the level of job satisfaction was problematic. Workers complained that they were alienated by the company's HRM practices. The unrest culminated in an illegal 26-day strike in October 2001 and the company dismissed a number of workers. After protracted negotiations, the company reinstated the majority of workers and implemented a series of HRM changes (Granjo, 2003). Employment increased from 2001 and by 2007/08 the Company was employing 1100 people across five different levels, including two categories of technician, administration staff, middle management and senior management.

The dynamics of job satisfaction are elusive because this phenomenon is influenced by such a wide range of explanatory variables (Bender et al., 2005; Faragher et al., 2005; D'Addioa et al., 2007). The formation of successful multicultural teams is also not a natural phenomenon (Rijamampianina & Carmichael, 2005) and the advancement of black employees in most African countries has been problematic (Human, 1996; Schneider & Barsoux, 2003c). The study this paper is based on investigated some of the HRM variables that influenced job satisfaction in the Mozal Company. Answers were sought to three research questions. The researchers examined the level of job satisfaction in a company setting, tested whether a range of intrinsic and traditional extrinsic HRM rewards influenced job satisfaction, and attempted to discover whether humanist variables influence job satisfaction in an African context. The paper is organised as follows: Section 2 describes the conceptual framework, Section 3 outlines the data and method, Section 4 presents and discusses the results, and Section 5 concludes and offers suggestions for future research.

2. A conceptual framework

The determinants of job satisfaction have been approached from a wide range of theoretical perspectives, including social psychology, neoclassical theory and management sociology (Mulinge & Mueller, 1998; Sousa-Poza & Sousa-Poza, 2000; Lévy-Garboua & Montmarquette, 2004; Diaz-Serrano & Vieira, 2005; D'Addioa et al., 2007). Explanations of ‘job satisfaction’ vary. It has been described as an exchange process, where people enter into work relationships in anticipation of rewards in exchange for their inputs (Kalleberg, 1977; Mulinge & Mueller, 1998; Van de Vliert & Janssen, 2002). Sousa-Poza & Sousa-Poza (2000) see it as an input–output function, where satisfaction results when the outputs (rewards) exceed the inputs (which might include education, working hours, effort, the physical conditions and the level of danger in the workplace). The rewards can be intrinsic or extrinsic. Intrinsic rewards include participation in decision making, autonomy, the facility to communicate, task significance, career growth and feedback on performance (Herzberg, 1966; Greenberg, 1980). Extrinsic rewards are more tangible and immediate benefits such as pay, fringe benefits, job security, working conditions and personal or group relationships (McGuire et al., 2005). A sub-category of these is labelled ‘convenience extrinsic rewards’, which include living a convenient distance from work, good working hours, pleasant working surroundings and an absence of job overload, role ambiguity and conflict (Kalleberg, 1977; Greenberg, 1980; Price & Mueller, 1986).

Other variables influencing job satisfaction could include health (Faragher et al., 2005), gender based variables (Bender et al., 2005; D'Addioa et al., 2007), education (Vila & García-Mora, 2005), age and length of service (Eskildsen et al., 2004; Lévy-Garboua & Montmarquette, 2004). Other possible influences are the level of the employee in the organisation (Huang & Van de Vliert, 2004), the absence of external opportunities, family responsibilities and firm-specific training (Mueller & Price, 1990; Mulinge & Mueller, 1998; Sousa-Poza & Sousa-Poza, 2000).

2.1 Cultural dimensions, humanism and job satisfaction

‘Culture’ can be variously defined. Schein describes it simply as ‘the deeper level of basic assumptions and beliefs that are shared by members of an organisation’ (1985:6). Rijamampianina offers a fuller definition, stating that:

Culture is created, acquired and/or learned, developed and passed on by a group of people, consciously or unconsciously, to subsequent generations. It includes everything that a group thinks, says, does and makes – its customs, ideas, mores, habits, traditions, language, and shared systems of attitudes and feelings – that help to create standards for people to coexist. (1996:124)

Clearly, one can anticipate that different cultures would prefer different kinds of HRM (Hofstede, 2001). Considerable evidence points to a collectivist and feminine society in Mozambique (Granjo, 2003; Webster & Wood, 2005; Webster et al., 2006) that is likely to prefer more humanist management practices (Hofstede, 2001; Jackson, 2002; Wooten & Crane, 2004). Collectivism is characteristic of ‘a society in which people from birth onwards are integrated into strong, cohesive in-groups’ (Voronov & Singer, 2002:260) and is an important factor in the Mozambican workplace. A feminine society is one in which ‘social gender roles overlap’, and this would also be extended into the workplace, creating the expectation that employees will be ‘modest, tender and concerned with the quality of life’ (Hofstede et al., 2009:297). In this type of cultural context, multinationals need to soften their instrumentalist type HRM programmes to accommodate the beliefs and feelings of Africans (Prinsloo, 2000; Inyang, 2009).

A traditional contract based view of employment should also be modified to incorporate an African type of humanism (Broodryk, 2006) that places great emphasis on the fact that employees are part of a collective. The implication of this is that the feelings and beliefs of the collective have a major influence on the individual (Nussbaum, 2003). HRM programmes that soften their instrumentalist position see employees as human beings rather than as resources and the firm adopts a father-figure role that is expected to be concerned with the employees' feelings and well-being. The firm is therefore expected to take care of employees, help them and be forgiving where genuine mistakes are made. This father-figure role is also expected to be extended to the welfare of employees' families and communities (Okpara & Wynn, 2008; Azolukwam & Perkins, 2009). A contract based view of employees, therefore, must be modified to recognise the importance of interpersonal relationships, personal dignity, and respect for others, particularly older (by implication, wiser) people (Kamoche et al., 2004; Wooten & Crane, 2004; McGuire et al., 2005). Collectively, these characteristics call attention to the need for humanism in the workplace (Wooten & Crane, 2004; McGuire et al., 2005) and humanist HRM programmes will develop work ideologies to link individuals to organisations by means of shared reasoning, norms and values (Wooten & Crane, 2004).

Note that in the context of this study the term ‘humanist’ means that management (the firm) are concerned about their employees as human beings, consider their feelings, and do not view them only in a work related context.

3. Data collection and methodology

The data collected during 2007 and 2008 were largely qualitative. A survey was combined with a case study and the data were analysed using a range of descriptive, regression and qualitative techniques. The reason for the case study methodology was to take advantage of the principle of triangulation and gather information about and examine a broad range of variables influencing job satisfaction. In view of the primarily qualitative, exploratory nature of the study, the results should be interpreted from the perspective of potential usefulness to multinationals investing in Africa, rather than as definitive or prescriptive (Jackson, 2002).

The questionnaire used for the survey was constructed on the basis of existing questionnaires that had been used by many other researchers to test for job satisfaction, the influence of rewards and the influence of humanist variables in the workplace (Mulinge & Mueller, 1998; Sousa-Poza & Sousa-Poza, 2000; Hofstede, 2001; Schneider & Barsoux, 2003a,b,c). The survey was facilitated and administered by the company management and the labour union during 2007. A random sample of 181 was collected after the human resources department and the labour union had administered a survey across five levels of employee. The sample size (n) of 181 out of 1100 (N) employees was split across five levels (Level 1, 47.5%; Level 2, 20.4%; Level 3, 8.3%; Level 4, 17.7%; Level 5, 6%). The sample size at each level was proportional to the total number of employees at that level.

To test the first research question, i.e. to measure the level of job satisfaction, the respondents were asked a single question: ‘Considering everything, how satisfied are you with your day-to-day job?’ The response to this question was measured on a six-point Likert scale, ranging from ‘completely satisfied’ (1) to ‘completely dissatisfied’ (6). Four measures of satisfaction were then aggregated to test the reliability of the single measure of job satisfaction. Descriptive statistics, combined with qualitative analysis, were used to analyse the data. The five levels of employees were then regressed against the dependent variable, job satisfaction, to determine whether there was any significant relationship between employee level and job satisfaction. A Kruskal–Wallis test was used to highlight the difference between groups with respect to their levels of job satisfaction across the five levels of employee.

To test the second research question, i.e. to discover whether a range of well-tested intrinsic and extrinsic rewards influence job satisfaction in an African context, a further 14 variables were examined. These were based on the Hofstede (2001) variables used to test an employee's perception of the ideal job in a survey involving 100 countries. Each of the variables was measured on a five-point Likert scale. Descriptive statistics, combined with some qualitative analysis, were used to analyse the data. To gain a clearer understanding of how the variables affected different levels of employee, the ranking of responses at employee level was also analysed.

To test the third research question, i.e. to discover whether humanist variables influence job satisfaction in an African context, 16 humanist variables were then selected from the job satisfaction literature of developing countries (Sousa-Poza & Sousa-Poza, 2000). Each of the variables required a response on a five-point Likert scale. Variables used to measure the level of corporate humanism were employee perception of tension, fear, pride, inspiration, caring-concern, support, community welfare and recognition of service rendered and employee age. To develop the dependent variable, namely the level of job satisfaction, four measures of job satisfaction were aggregated in order to develop a more comprehensive measurement of job satisfaction that included elements of humanism. Ordered logistic regression (univariate and multivariate) using STATA was then used to analyse the relationship between job satisfaction and each of the 16 predictor variables.

In line with the exploratory approach, a case study method was used to collect further data from various stakeholders in the organisation. This method allowed the researchers to investigate and interpret a wider range of variables that could influence job satisfaction in a company context, and to access additional data, including company documents and media reports, and review the history of the company from its formation (Leedy, 1993; Yin, 1994; Leedy & Ormrod, 2001). The data were collected using semi-structured in-depth interviews with two shop stewards, the paymaster, the HR manager, the HR superintendent, the industrial relations superintendent, the internal audit controller, a Level 4 manager, a Level 5 manager and the finance manager. Additional interviewees were a lawyer, a consultant and the chairman of the labour union. The data collected from the interviews were written up and subjected to thematic content analysis before being reviewed by the company.

The study had a number of limitations. First, its exploratory nature means that it raises more new questions to be investigated than it answers the ones it poses. Second, the case-based results do not lend themselves to generalising any of the conclusions.

4. Results and discussion

This section describes the findings in answer to the three research questions: the level of job satisfaction, the importance of intrinsic and traditional extrinsic HRM rewards, and the influence of a series of humanist variables on job satisfaction.

4.1 The level of job satisfaction

Table 1 shows the level of job satisfaction in the Mozal Company as compared with the results obtained from 21 countries by Sousa-Poza & Sousa-Poza (2000). Using the six-point Likert scale described above, the results showed that, of all the countries in the table, Mozal appeared to have the lowest score for the category of ‘completely satisfied’ and the second lowest score for ‘very satisfied’. Conversely, it had the fourth highest score for ‘fairly satisfied’ and the second highest score for ‘neither satisfied nor dissatisfied’. This indicated that the worker's level of job satisfaction in this company was most comparable with that of a series of former Soviet Bloc countries (Eastern Europe) that have also encouraged foreign based investment in recent decades (Sousa-Poza & Sousa-Poza, 2000). An alternative measure of job satisfaction, that involved aggregating the four measures of job satisfaction, appears to support the single measure used (see figures in parentheses in Table 1).

Further analysis of the company job satisfaction data corroborates Huang & Van de Vliert's observation (2004) that job satisfaction can be influenced by the level of the employee in the organisation. The results of the present study showed that as the employee level increased from Level 1 (unionised factory workers) to Level 5 (management) the level of job satisfaction increased. The results of an ordinal regression indicated a highly significant relationship (less than 1%) between job satisfaction and employee level. A Kruskal–Wallis test further confirmed that there was a significant difference in job satisfaction across the five levels of employee. The chi-square score was 15.629 (4 degrees of freedom) and the probability that there was no difference between the levels of satisfaction across the five levels was 0.0036, which is less than 1%. The differential level of satisfaction, further illustrated in Figure 1, clearly demonstrated that Level 5 employees (management) had significantly higher levels of job satisfaction than the other four levels. (Four measures of job satisfaction were aggregated, hence a score of 5 indicates a very high level of job satisfaction and 30 a very low one.)

Table 1: Level of job satisfaction – descriptive statistics

Country	Completely satisfied (%)	Very satisfied (%)	Fairly satisfied (%)	Neither… nor (%)	Fairly dissatisfied (%)	Very dissatisfied (%)
Mozal survey (n = 181) (Qs 69–72)*	3.8 (4.9)	14.4 (16.9)	43.5 (39.4)	27.0 (26.9)	7.5 (8.0)	3.8 (3.7)
Anglo Saxon
United States	15.9	32.5	35.0	7.3	6.4	2.9
Great Britain	13.0	22.6	43.3	10.3	8.1	2.7
New Zealand	11.8	29.4	42.6	9.1	5.2	1.9
Western European
Denmark	24.7	36.9	26.3	7.7	3.6	0.8
France	11.4	22.3	42.2	14.1	7.7	2.3
Germany	7.9	28.6	43.3	14.5	4.3	1.4
Italy	15.5	19.9	43.4	13.1	5.0	3.1
The Netherlands	10.5	37.6	41.8	6.1	2.8	1.2
Norway	10.1	26.6	45.4	12.3	4.6	1.0
Portugal	17.7	21.2	32.5	21.8	3.7	3.1
Spain	13.8	36.0	33.7	11.3	3.7	1.5
Sweden	9.9	30.4	42.1	11.7	4.2	1.7
Switzerland	11.7	40.3	37.1	7.2	2.4	1.3
Eastern European
Hungary	9.1	14.1	39.0	29.6	5.0	3.2
Czech Republic	15.4	13.0	50.5	14.3	4.7	2.1
Slovenia	11.6	15.4	39.7	26.2	5.2	1.9
Bulgaria	10.7	18.4	49.3	13.0	4.8	3.8
Russia	13.2	26.7	19.2	25.5	8.0	7.4
Others
Japan	5.5	24.7	41.6	13.4	9.7	5.1
Israel	27.6	20.8	32.8	10.9	4.9	3.0
Cyprus	21.6	38.2	27.4	10.3	2.0	0.5
Total	13.6	26.5	38.6	13.8	5.1	2.8
*These were the question numbers on the questionnaire for the values in parentheses. The level of job satisfaction was also tested by aggregating four different measures to provide a broader alternative.
Source: International figures from Sousa-Poza & Sousa-Poza (2000).

Figure 1: Job satisfaction and employee level

The interviews provided some further insights. In 2007, a manager (Samuel) said the lowest level of satisfaction was demonstrated by Level 1 and Level 2 employers (who were the principal backers of the 2001 strike). This statement was supported by a Level 2 maintenance technician (Gafar) who said that ‘the gap between Level 1 and Level 2 and management is too big, but Level 4 and 5 are now also feeling their situation is compromised because they are not unionised and the salary gap between us is getting smaller’.

4.2 The influence of rewards on job satisfaction

Hofstede's 14 intrinsic and extrinsic rewards (2001) were all ranked as important determinants of job satisfaction (see Table 2). Although ‘challenging work’ has been rated as a top retention and motivation strategy in some high technology industries (Horwitz et al., 2003), it was somewhat surprising to see that the respondents in this study rated this reward, and also ‘good training opportunities’, more highly than ‘opportunity for high earnings’ and ‘good fringe benefits’, given the central role of these latter two rewards as a precursor to the strike of 2001. The importance of job quality, however, is widely cited in the literature (Herzberg, 1966; Kalleberg, 1977; Mulinge & Mueller, 1998; Pfeffer, 1998; Sousa-Poza & Sousa-Poza, 2000; Hofstede, 2001; Van de Vliert & Janssen, 2002). The respondents ranked ‘level of cooperation with others’, a social extrinsic reward, in second place, which underlines the importance of in-group relationships in a collective workforce (Hofstede, 1994).

Table 2: Rewards influencing the perception of the ideal job – descriptive statistics (n = 181)

			–Binomial exact –		Means and ranking
Reward*	Obs.	Median	[95% confidence interval]		Mean	Rank
Challenging work	170	2	1	2	1.77	1
Level of cooperation with others	176	2	1	2	1.81	2
Good training opportunities	177	2	1	2	1.86	3
Job security	176	2	1	2	1.93	4
Good working relationship with manager	177	2	1	2	1.94	5
Fully use skills	172	2	2	2	1.95	6
Live in desirable area	177	2	1	2	1.95	7
Good physical working conditions	179	2	1	2	1.96	8
Opportunity for high earnings	171	2	1	2	1.96	9
Sufficient leisure time	176	2	2	2	1.98	10
Opportunity for advancement	172	2	2	2	2.01	11
Recognition for good work	177	2	2	2	2.03	12
Good fringe benefits	175	2	2	2	2.07	13
Freedom to adopt own approach	175	2	2	2	2.07	14
*Rewards based on Hofstede (2001).

‘Job security’ was also highly ranked. In the context of limited alternative employment opportunities, as well as the responsibility for an extended family, job security is a crucial household variable in Mozambique (Granjo, 2003; Davis, 2008). Another important finding was the low ranking of ‘good working relationship with manager’ – the negative impact of a poor relationship with management is a major source of grievances. A collective workforce, in particular, expects a manager to be ‘one of them’, and to use intuition and be able to understand their feelings (Hofstede, 2001). Lower ranked rewards indicated the descending importance of self-development (‘training opportunities’), good physical working conditions, earnings and leisure time. Finally, the respondents ranked recognition, fringe benefits and a degree of autonomy in the workplace, as their lowest ranked (but still important) rewards.

Job titles were cited as being of the utmost importance by nearly every one of the interviewees, in support of the findings of Granjo (2003).[0] Job titles were particularly important for maintenance personnel, who did not like to be called ‘maintainers’ but rather ‘maintenance technicians’.

To refine the analysis of the responses, the variables influencing the ideal job were analysed across the five levels of respondent. The results in Table 3 show that the five levels of employees ranked some of the variables very differently. However, every category of employee except Level 2 (maintainers) ranked ‘good training opportunities’ as one of the top six rewards. Level 2 employees felt that the training they had received should have resulted in the same pay as expatriates. They were also unhappy that they were called ‘maintainers’ and not ‘maintenance technicians’ (Gafar, maintainer; Dumi, HR manager; Mark, HR superintendent; Tomas, industrial relations superintendent).

Table 3: Ranking of ideal job variables – descriptive statistics (n = 181)

	Level 1	Level 2	Level 3	Level 4	Level 5
Challenging work	2	6	2	1	1
Level of cooperation with others	1	5	1	4	5
Good training opportunities	5	11	4	5	6
Job security	4	7	11	2	13
Good working relationship with manager	6	8	6	8	7
Fully use skills	13	13	3	6	4
Live in desirable area	8	2	9	9	11
Good physical working conditions	3	12	13	3	8
Opportunity for high earnings	10	1	8	13	2
Sufficient leisure time	9	4	7	12	14
Opportunity for advancement	12	10	5	7	9
Recognition for good work	7	14	10	10	10
Good fringe benefits	14	3	12	14	3
Freedom to adopt own approach	11	9	14	11	12

The importance of physical conditions was especially highly ranked by Level 1 operators. They felt that the working conditions were exhausting because they had had no similar experience in their working life, and some felt that they were dangerous (Brief & Weiss, 2002; Granjo, 2003). JJ Muhai (an internal audit controller) said his family members could not understand why he was exhausted at the end of a day's work, and Mulayo (the paymaster) said that many workers felt there was danger because they remembered how dangerous it had been when their families worked in South African mines.

Although the survey ranked many of the rewards higher than earnings and fringe benefits, the case study interviews contradicted these findings. Four interviewees offered the opinion that a prime reason for the strike of 2001 was that Levels 1 and 2 thought they were underpaid relative to the training they received, and these four also said that Mozambican employees should be paid the same as their expatriate colleagues (Samuel, manager; Dumi, HR manager; Mark, HR superintendent; Tomas, industrial relations superintendent). Two interviewees pointed out that even management was now complaining that the salary gap between them and the workers was being unfairly reduced because they were not unionised (Elias, chairman of labour union; Gafar, maintainer), and three observed that fringe benefits such as food allowance, provident fund, study assistance and particularly medical aid and programmes for HIV/AIDS were very important (Elias, chairman of labour union; Gafar, maintainer; JJ Muhai, internal audit controller).

4.3 How do humanist variables influence job satisfaction?

To test the third research question the humanist variables listed in Table 4 were individually regressed on job satisfaction. The results highlighted some interesting features of the workplace. Quite surprisingly, the level of tension in the job situation (TENSE) and the management style of the immediate superior (MANAGER) did not appear to influence job satisfaction significantly, which contradicts the findings of Hofstede (2001). Another finding that contradicts Hofstede's is that job satisfaction was negatively influenced because workers were afraid to disagree with their managers (AFRAID p < 10%), and that it was important to have a good relationship with the manager. However, Mulinge & Mueller (1998) suggest that in some collectivist countries the ability to communicate and be heard by management is more important than the management style, and Granjo (2003) suggests that workers in Mozambique may accept autocracy and power inequality in the workplace as a fact of life. Questioning the judgement of a superior could be taken as offensive and lead to conflict, which would disrupt the harmony of the workplace (Nussbaum, 2003; Kamoche et al., 2004; Broodryk, 2006).

Table 4: Humanist variables influencing job satisfaction – ordered logistic regression (univariate)

Predictor	Odds ratio	Std err.	Z	p > z	[95% confidence interval]		Significance
TENSE	–0.2215232	0.2117084	–1.05	0.295	–0.636464	0.1934176
INTENTION	–0.5180336	0.1972114	–2.63	0.009	–0.9045609	–0.1315064	*
AFRAID	–0.3162494	0.1709976	–1.85	0.064	–0.6513985	0.0188996	**
MANAGER	0.0651599	0.1485009	0.44	0.661	–0.2258965	0.3562164
EFFORT	0.2517099	0.1717664	1.47	0.143	–0.084946	0.5883658
PRIDE	0.8230633	0.1684091	4.89	0.000	0.4929876	1.153139	*
INSPIRESME	0.9702437	0.2164708	4.48	0.000	0.5459688	1.394519	*
VALUES	0.4892577	0.1712613	2.86	0.004	0.1535918	0.8249237	*
MOZCARES 1	0.6527916	0.1623787	4.02	0.000	0.3345352	0.971048	*
MOZCARES 2	0.9860946	0.1960673	5.03	0.000	0.6018097	1.37038	*
HELPSME	0.6866854	0.1619803	4.24	0.000	0.3692099	1.004161	*
FORGIVESME	0.4633704	0.1491115	3.11	0.002	0.1711172	0.7556236	*
COMMUNITY	0.4027492	0.1769409	2.28	0.023	0.0559514	0.7495471	*
ENVIRONMENT	0.1240686	0.170295	0.73	0.466	–0.2097035	0.4578408
SERVICE	–0.3230725	0.1890899	–1.71	0.088	–0.6936818	0.0475368	**
AGE	–0.4894568	0.1713452	–2.86	0.004	–0.8252872	–0.1536265	*
*Significant at the 5% level. **Significant at the 10% level.

Other variables that had a significant positive influence on job satisfaction were the respondents' level of pride in the company (PRIDE), being inspired by the company (INSPIRESME), their feeling that the company's values were similar to theirs (VALUES) and the perception that Mozal cared for their well-being (MOZCARES 1). Finally, additional factors positively influencing job satisfaction were the perceptions that the company cared about their opinions (MOZCARES 2), helped them as employees (HELPSME), and had invested in their local communities (COMMUNITY). Finally, the age of employees (AGE) was significantly negatively correlated with job satisfaction.

The significant variables, identified in the univariate analysis, were then used as the basis for constructing a multivariate model (see Table 5). In the multivariate results only two variables emerged as significant. This further analysis showed that job satisfaction was significantly positively influenced as a result of workers being inspired (INSPIRESME) by the company (p < 10%). There also appeared to be a highly significant negative relationship between the age (AGE) of employees and job satisfaction (p < 5%). This confirms findings from other collectivist African countries where age is negatively correlated with job satisfaction (Eskildsen et al., 2004) and older employees often feel their status in the company is not properly recognised (Nussbaum, 2003; Kamoche et al., 2004).

Table 5: Humanist variables influencing job satisfaction – ordered logistic regression (multivariate)

Predictor	Coefficient	Std err.	z	p > z	[95% confidence interval]		Significance
INTENTION	–0.3638956	0.2293895	–1.59	0.113	–0.8134908	0.0856997
AFRAID	–0.0585321	0.2039804	–0.29	0.774	–0.4583264	0.3412622
PRIDE	0.2212715	0.2530603	0.87	0.382	–0.2747177	0.7172606
INSPIRESME	0.6006747	0.3112362	1.93	0.054	–0.0093371	1.210687	**
VALUES	–0.170849	0.2458507	–0.69	0.487	–0.6527076	0.3110096
MOZCARES 1	0.0602553	0.2507922	0.24	0.810	–0.4312884	0.551799
MOZCARES 2	0.3676478	0.3316529	1.11	0.268	–0.2823799	1.017676
HELPSME	0.2366562	0.2257508	1.05	0.294	–0.2058073	0.6791197
FORGIVESME	0.1396051	0.2023747	0.69	0.490	–0.2570421	0.5362522
COMMUNITY	–0.1456978	0.2643735	–0.55	0.582	–0.6638604	0.3724648
SERVICE	–0.2978317	0.2315025	–1.29	0.198	–0.7515683	0.155905
AGE	–0.4774027	0.2185087	–2.18	0.029	–0.9056718	–0.0491335	*
*Significant at the 5% level. **Significant at the 10% level.
Notes: Ordered logistic regression. Number of obs = 108; LR chi2 (12) = 51.49; Prob > chi2 = 0.0000; Log likelihood = –80.672879; Pseudo R ^2= 0.2419.

The interviews, conducted during 2007, revealed some additional humanist variables that were strongly responsible for lower levels of job satisfaction before the strike of 2001.

Nine interviewees expressed the opinion that management were ‘not human’ before the strike in 2001, but were ‘aggressive’, ‘military’ and ‘racist’, would only talk but not listen, and workers were scared (Daniel, lawyer; JJ Muhai, internal audit controller; Gafar, maintainer; Riaan, finance manager; Enrique, superintendent; Dumi, HR manager; Mark, HR superintendent; Tomas, industrial relations superintendent; Lino, paymaster). Seven said that the Company was now communicating with its staff, sharing its concerns with them, using the public and internal communications to cover news such as family births and deaths, and on the whole seeing workers ‘more as humans’ (Samuel, manager; Dumi, HR manager; Mark, HR superintendent; Tomas, industrial relations superintendent; JJ Muhai, internal audit controller; Gafar, maintainer; Lino, paymaster). And, finally, one said that despite these changes many workers still felt the workplace was over-regimented and over-structured (Daniel, lawyer).

The results demonstrate the strength of the employees' feelings about the importance of humanism in the workplace (Kamoche et al., 2004). When job satisfaction was measured using four aggregated measures, one of the questions asked was ‘whether their families were satisfied with Mozal as an employer’. Interestingly, the respondents' level of job satisfaction was identical to their families' perceptions of the company, indicating the importance of the role of family in a developing collectivist country (Mulinge & Mueller, 1998; Granjo, 2003; Carmichael & Rijamampianina, 2007; Dimba & K'obonyo, 2007; Horwitz, 2007).

5. Conclusion and recommendations

The paper has illustrated some of the difficulties that confront multinationals when they implement HRM programmes in a developing African country. The lesson that has been learned is that local employees are at least as motivated by traditional extrinsic rewards as their developed country counterparts – both our study and the literature support this observation They want, however, to earn these rewards in a more humanist environment, where caring about people is visibly demonstrated. Furthermore, they are less concerned about rules than about the way the rules are administered. The case study interviews also revealed that none of the technical standards and production procedures (for the production of aluminium) had been changed – they could not be or the product would not be of the necessary quality and the production environment would be too dangerous.

The level of job satisfaction showed that there were some similarities with former Soviet Bloc countries that have also experienced foreign based HRM programmes that were superimposed on their predominantly collectivist local workforces. The complex nature of job satisfaction was also underlined by significant differences across employee levels. A combination of these insights suggests that company HRM programmes need to be targeted to specific levels in an organisation, and to accommodate the requirements of the local culture.

The results support the contention that the traditional extrinsic and intrinsic rewards that form the basis of instrumentalist HRM are equally important motivators of job satisfaction in an African context. The Mozal Company, the focus of this study, responded to the challenge of the strike in 2001 by completely re-investigating their employment package. The findings showed that to achieve job satisfaction in its workforce it was important for the company not only to modify the rewards but also to adopt a more humanist management style. The approach taken by South African Breweries (SAB) when they partnered with a local brewery in China as part of their internationalisation strategy is an example of the extent to which organisational success rests on implementing locally adapted strategies – and a way of doing it successfully (Vermeulen, 2001). The key element of the SAB success lay in the fact that they absorbed and identified key cultural elements and ways of going about business before implementing HRM and other practices. Humanist variables, therefore, appeared to mediate the effectiveness of traditional rewards, as well as to support a central tenet of African tradition, namely the concept of ubuntu or ‘humanness’, which stresses that a ‘person is a person through other people’. The improvement in job satisfaction after the 2001 strike, therefore, was due as much to the change in management's attitude to and communication with the employees as to the revised compensation package. A major lesson that was learned was that management was required to understand the culture and feelings of their workers better, in much the way that Vermeulen (2001) describes. Finally, an important challenge identified in the study was the status attached to age in the workplace and the need reconcile this with added value.

Many of the variables used in this study to test the importance of humanism in the African workplace could be more vigorously investigated across a wide range of countries and technological conditions, and the humanist variables should be tested against the tenets of the African tradition of ubuntu in order to examine the possibility of developing specifically African HRM programmes.