ARTICLE HEALTH AND POLICE UNIONISM IN THE EASTERN CAPE: ATTITUDES, PERCEPTIONS AND PROSPECTS1 Geoffrey Wood As the employer of over a million South Africans, the public sector provides jobs to a large segment of South Africa's economically active population. Dominated by whites, for many years this sector was known for extremely conservative employment practices, and unequal treatment for members of different racial and ethnic groups. Until recently, most employee associations reflected this reality, being largely white dominated and with a tendency to avoid confrontations with senior administration. This picture changed dramatically in the late-1980s, with the emergence and rapid expansion of independent unions in the public sector. This paper focuses on the nature and impact of public sector unionism involving hospital and police workers in two urban centres in the Eastern Cape Province, South Africa. These two divisions were selected as a result of their relative importance within the public sector as a whole, constituting 'essential services', and on account of the existence of rapidly expanding unions therein. By 1992, nurses represented some ten per cent of public sector employees, members of the armed services 21 per cent, and labourers (many employed by the police or hospitals) 22 per cent (SALB 1992:48-9). In contrast to police and hospital workers, those employed in other areas of the public sector, with the possible exception of teachers, seem somewhat less inclined to engage in collective action, periodic threats notwithstanding. This study is based in the Eastern Cape. It is appreciated that some of the trends and practices noted may not be true on a national level, although the specific dynamics may reveal some new light into the internal operations of South Africa's new public sector unions. TRANSFORMATION 33 (1997) ARTICLE WOOD Organising Public Sector Workers: The Case of Police and Health Workers Founded in the late-1980s by former police lieutenant, Gregory Rockman, the Police, Prisons and Civil Rights Union (POPCRU) had a total membership of 5000 by 1990, and has expanded rapidly since then. Its initial membership drive involved visiting police stations, prisons and community organisations, informing workers of the union (Mtshelwane 1993:70). Originally, the union had its strongest organisational presence in the prisons, but since then organisation of members of the police has proceeded apace. During 1993, the Correctional Services offered to recognise POPCRU, but that union instead pushed for a blanket recognition agreement from both police and the prisons. Although POPCRU has formally decided to affiliate to COSATU, this decision has yet to be implemented. POPCRU has experienced serious financial difficulties, probably as a result of internal mismanagement. POPCRU does not have conventional shopfloor structure, and is structured as per the following organogram: ^National Executive Committee') Branch) {Branch) (Branch) (Branch In the early-1990s, an alternative police union emerged, in the form of the South African Police Union (SAPU). SAPU was initially seen by many as an attempt by right wing elements to establish a conservative counterweight to POPCRU. However, since its establishment SAPU has adopted an increasingly critical stance, and has amassed a particularly strong support base in the Western Cape. Today, SAPU is a predominantly black trade union, with a larger claimed membership than POPCRU. Nonetheless, it remains deeply divided between conservative and radical elements, and there is evidence that, as is the case with POPCRU, it has financial difficulties (see Mail & Guardian February 2,1995). SAPU has repeatedly called for 'decent' salaries, and permanent employee status for assistant constables (Argus June 29, 1994). SAPU members claim that they TRANSFORMATION 33 (19Q7) WOOD ARTICLE have been victimised by those in senior ranks, and, inter alia, have on occasion been placed on standby duty to prevent them attending union meetings (Argus June 29, 1994). A vocal minority within SAPU, led by Colonel Kuipe have continued to press for affiliation to COSATU (Cape Times March 31, 1994). However, SAPU remains opposed to using the strike weapon (Cape Times March 31,1994). Together, the two unions represent over 57 000 members of the police services (Plainsman June 29, 1994). Drawing together workers formerly represented by a range of COSATU-affiliated unions ranging from South African Allied Workers Union (SAAWU) to the Transport and General Workers Union (TGWU), National Education and Health Workers Union (NEHAWU) was launched in January 1987. Although intended to be COSATU's umbrella union in the public sector, it was decided that on account of a relatively small membership, it would initially focus on health and education workers (Baskin 1991:198). In its first year of existence, NEHAWU's membership actually dropped from over 9000 to 5800. However, as Baskin (1991:198) notes, this was probably more due to improved organisational procedures than to an actual decline. By the 1990s, NEHAWU began to experience rapid growth, particularly amongst health workers. However, the considerable diversity of employment categories was to pose a major challenge to NEHAWU. Whilst many nurses joined the 1992 health workers' strike organised by that union, it was evident that many also had conflicting loyalties (Keet 1992:51). In particular, it faced opposition from the South African Nursing Association (SANA). With a predominantly white, and traditionally conservative leadership, the Association has focused on safeguarding the professional status of nursing, yet, in the past, it gave somewhat limited attention to the issue of wages and working conditions. In reaction to this, well over 10 000 nurses joined NEHAWU since 1982 (Keet 1992:51). However, tensions between nurses and service workers have led to divisions within that union, and the emergence of a number of splinter unions specifically for nurses, most notable the Democratic Association of Nurses of South Africa, and the National Union of Nurses of South Africa (NUNS A) (Keet 1992:53). In addition, SANA has begun the process of reorganisation, which includes the establishment of a shopfloor organisation, and to combine the functions of a professional organisation with that of a trade union. The other major union operating in the health sector is the Public Servants' League (the PSL, not to be confused with the Public Servants' Association2) which describes itself as a 'non racial union ... dedicated to organising public sector workers, general working class unity, democratic shopfloor structures, healthy and safe working conditions' (SALDRU 1991:343). Recognised by the TRANSFORMATION 33 (1997) ARTICLE WOOD Commission for Administration as a Personnel Organisation, the PSL's membership increased rapidly in the 1980s, from 24 000 in 1983 to 46 000 in 1990 (SALDRU 1991:342). A relative newcomer to the union scene, another union, Hospersa, has established a presence particularly amongst supervisory and clerical workers. It has already clashed publicly with NEHAWU on key policy issues, and has threatened to embark on a major strike in the Eastern Cape region over demands for an improvement in service conditions of its members. Method The two case studies are based on complementary sources: two parallel questionnaire surveys, and a limited number of open-ended in-depth interview studies, the former of individual workers, and the latter of union officials. The hospital survey was based on a stratified random sample of hospitals in Port Elizabeth and Grahamstown3. The interviews of police were conducted in the Albany region4, and no problems of access were encountered . The sampling methods employed ensured a reasonable cross-section of workers in both cases . It should be noted that this research, by focusing on two groupings within the public sector, in a specific region, only provides limited insights into the attitudes held by members of the sector as a whole. Nonetheless, it provides new information on extent of union penetration, on divisions between unions, workers' expectations of unions, and their perceptions of the attitudes of senior administration thereto7. Findings: The New Challenge Police Unionism: Divisions and Hopes The majority of police service members encompassed by the survey were unionised. On the one hand, it should be noted that the process of community policing is relatively well-advanced in the area where the survey was conducted, which may have made senior administration more receptive to unionisation initiatives. On the other, it should be noted that the Eastern Cape was one of the main areas of POPCRU activity in the early days of that union. In-depth interviews of members of the police service revealed that the degree of unionisation in the Correctional Services is very much higher, with the majority of warders being members of POPCRU. By a small margin, the majority of unionised members surveyed belonged to POPCRU, although on a national basis it seems that the converse may be true. However, almost as many belonged to SAPU. It would be somewhat over-simplistic to ascribe the rapid expansion of the latter union simply as a conspiracy of senior administration. Whilst there have TRANSFORMATION 33 (1997) WOOD ARTICLE been persistent allegations that S APU originated as a result of a security police initiative8, it has since shed this conservative image, whilst it is evident that that union must have been seen to have offered the prospect of real benefits to its members to account for its rapid growth. A small minority of police service members belong to the traditionally exclusivist Public Servants' Association (PSA). Unlike POPCRU and SAPU, this union's membership is not confined to the Police and Correctional Services. The mode9 of respondents supported union unity. Interestingly, many SAPU members favoured unity with POPCRU and vice versa. It is evident that many do not see the two philosophies of unionism as mutually exclusive, as reflected by the desire of sectors of SAPU to affiliate to the COSATU federation, and that both unions incorporate more conservative and militant elements. However, POPCRU's leadership is far more politically outspoken, and, unlike SAPU, that union has a record of engaging in strike action. Although many supported union unity, a sizable number of respondents (29 per cent) were uncertain as to whether union unity was desirable. It is evident that the relatively new phenomenon of unionisation has left many in the police service somewhat uncertain as to what strategies the unions should adopt. It is therefore unlikely that a single police union will emerge in the foreseeable future. Whilst this division may allow the central government a greater freedom of action, it is indeed possible that it might face two contradictory pressures, to restructure the police, yet provide reassurance to those already occupying the junior and middle ranks of the service, potentially alienating both groupings. Respondents were then asked whether they believed senior administration saw unions as opposition. The mode of respondents (30.2 per cent) were uncertain as to whether this was the case. Roughly the same number of respondents believed senior administration saw unions as opposition as thought this was not the case. Again, it is indicative that many remain uncertain as to the implications of unionisation, and the responses of senior administration thereto. Interestingly, a large proportion of both SAPU and POPCRU members were uncertain of this -it is evident that if the former union has a 'special' relationship to senior administration, this is not at all evident to the rank-and-file in terms of their everyday experiences. In other words, the membership of both unions see their respective organisations as operating independently of senior administration, and as potentially capable of challenging it, and thus the central government. The mode of respondents (43.8 per cent) believed that trade unions should have the right to be represented at meetings of senior administration, whilst, again, a large minority of respondents were uncertain as to whether this was desirable. The relationship between union membership and attitudes towards TRANSFORMATION 33 (1997) ARTICLE WOOD representation at meetings of senior administration is depicted in Table 1 (percentage responses): Table 1 Favour representation N/R Yes Unsure No Totals at meetings of senior administration Union membership N/R 80 20 0 0 3 POPCRU 14 58.1 25.6 2.3 25.4 SAPU 0 60 30 10 23.7 PSA 0 28.8 42.9 28.6 4.1 Not a union member 8.1 20.7 36.5 25.7 43.8 Totals 8.5 43.8 31.4 15.4 100 As is evident from the above table, there is again no significant difference in attitudes towards representation at meetings of senior administration and union affiliation. Roughly 60 per cent of members of both major police service unions believed this to be desirable. Similarly, roughly the same proportion were undecided, or chose not to answer the question (perhaps a similar measure of uncertainty). It was amongst PSA members, and the unionised that a sizable proportion of respondents were opposed to union representation. The PSA has always had a different, more exclusivist philosophy to unionism, but does not seem to have been able to capitalise on the wave of unionisation in the police service to the same extent as the other two unions. Interestingly, large numbers of the un-unionised, as well as PSA members, were undecided on this issue. The mode of respondents believed the police should not have the right to strike (42 per cent). However, a sizable minority were in favour of this (34.3 per cent). More than any other issue, this reveals the differences in strategy between the two major police unions. A considerably larger number of POPCRU respondents supported the right to strike. However, a minority of POPCRU members were opposed to striking, whilst a minority of SAPU members favoured the use of this form of collective action. By March 1995, SAPU had yet to officially make use of the strike weapon, although it has engaged in go-slow action. In contrast, the initial POPCRU recruitment drives often went hand-in-hand with instances of collective action over long standing grievances. On the one hand, it is evident that fewer police service members are inclined to engage in collective action than workers in other areas of the state (see below). TRANSFORMATION 33 (1987) WOOD ARTICLE It would seem that, in the short term at least, the state would most likely face sustained challenges from other areas of the public sector. On the other hand, strike action by POPCRU members could potentially be politically explosive in that that union represents employees occupying the lowest occupational category of police assistant, the overwhelming majority of which are Africans. Such strikes would be particularly likely if it appeared that the promotion of blacks in large numbers into the middle ranks of the service appeared slow. As noted elsewhere, there are certain minimum educational requirements that preclude many police assistants from joining the regular force. In other words, there is no automatic career path for police assistants. In addition, it should be noted that there have been a number of strikes involving police members in the Transkei, who are opposed to the introduction of the present criteria for promotion operative in South Africa. Table 2 compares union membership with rank held. In other words, it examines the relationship between relative seniority and allegiances to the union movement. Tablo 2 Union N/R POPCR U SAPU PSA Not a Totals union member Rank Major - Colonel 0 0 0 20 80 3 l_t - Captain 0 0 28.6 0 71.4 4.1 Warrant Officer 0 12.5 37.5 0 50 4.7 Sergeant 3.3 20 26.7 10 40 17.8 Constable 1.6 23 41 0 34.4 36.1 Police Assistant 7.1 64.3 7.1 0 21.4 8.3 Totals 3 25.4 23.7 4.1 43.8 100 It is evident that the only union to have significantly penetrated the more senior ranks (major and upward) is the conservative PSA. As noted earlier, PSA members seem to have rather different attitudes on key issues than the other two unions. The PSA has not enjoyed the same rapid rate of expansion within the service as the other two unions. However, the senior position occupied by many of its members implies that its capacity to challenge central government cannot be underestimated. SAPU has made important inroads into the ranks of the junior officers, and, in addition, organised over one third of Warrant Officers. In contrast, the most senior POPCRU members surveyed were Warrant Officers TRANSFORMATION 33 (1997) ARTICLE VVOOD (12.5 per cent of all holding this rank). In other words, POPCRU members were concentrated in the lowest ranks. It should be noted that there remains a statistically significant linkage between race and rank status within the service10. Africans are overwhelmingly clustered in the most junior ranks in the service, and form the vast majority of police assistants. The highest ranking African encompassed in the sample was a Warrant Officer. However, closer examination revealed that the mode of Constables (41 per cent) were SAPU members, whilst POPCRU accounted for only 23 per cent. In contrast, over 64 per cent of Police Assistants (sometimes known as 'kitskonstabels') were POPCRU members, compared to only 7.1 per cent for SAPU. In other words, POPCRU has achieved the greatest degree of penetration amongst the most marginalised, members of the service that do not have the same status, training, or security of tenure as the regular police. In addition, police assistants often bore the brunt of community protests during the apartheid era. Membership of POPCRU would give members not only a collective voice, but also as a result of that union's political alliances, a greater effective legitimacy within the community. It should be noted that POPCRU also has a strong following amongst prison warders". Both POPCRU and SAPU have significant follows amongst Africans in the regular police. This, then is the major division between the two main police unions, SAPU representing often more senior members of longer standing, and POPCRU the more junior ranks, especially those of only assistant level. The minimum educational requirements for admission to the service as police assistants are far lower than for the regular force. A greater emphasis on recruiting blacks into the service will probably not result in the lowering of the educational requirements, which would open the regular force to police assistants. In other words! affirmative action programmes will neither necessarily reduce the divisions between the assistants and the regular, nor increase the possibility of upward mobility for the former. As noted above, SAPU seems to have had considerable success in recruiting black members of the regular force. Whilst POPCRU's capacity to expand into the more senior ranks should not be underestimated, it seems evident that one of the major reasons for divisions between the two police unions for the foreseeable future will remain the distinction between the regular force and the assistants. It generally tends to be the case that those groupings who have the fewest promotional prospects, the least desirable conditions of service, and lower prestige/professional status will tend to favour the more militant forms of collective action. It should be recognised that the regular force and police TRANSFORMATION 33 (1987) WOOD ARTICLE assistants may have different concerns and, again, that the state may thus face contradictory pressures from the two unions. The mode of respondents (49.7 per cent) believed that union leaderships were accountable to their memberships. In other words, almost one half of respondents believed that unions accurately represented their wishes. A sizable grouping were unsure (32.5 per cent) as to whether this was really was the case, but this was largely composed of the ununionised, who would have limited experience of participation in union activities. There did not seem to be a linkage between union membership and responses to this question. In other words, to their respective members, S APU did not seem any more accountable than POPCRU or vice versa. On the one hand, it is probably the case that SAPU's less confrontational stance would be more acceptable to many in senior administration. On the other, it is evident that it does have a definite constituency, and, as such, reflects the demands and concerns of a significant sector of the police service. Most respondents were either uncertain, or believed there was little co-operation between senior administration and their union. Only 16.6 per cent of respondents believed this to be the case. In other words, whilst the process of unionisation is comparatively advanced, the bargaining process has yet to become institutionalised, or formalised. This may make both police unions more willing to mount a challenge to the state in future. Similarly, few respondents believed that senior administration saw their union as a means of effective management. Meanwhile, only 30.2 per cent of respondents believed that communication channels within the police were currently adequate. Indeed, there are none of the formalised collective bargaining structures present in the private sector, albeit that legislative changes may redress this. In summary, it can be seen that a number of general trends are apparent in police unionism. Firstly, it is evident that the two major unions, SAPU and POPCRU, have significant, but different constituencies. The former is more popular amongst longer standing police service members, and the latter, in addition to its power-base in the prisons, amongst the junior ranks, especially police constables. Although a sizable proportion of the memberships of both unions support union unity, with seemingly much common ground, there are certain important differences which make a merger in the short term unlikely. These differences revolve around issues such as the differing status between regular police service members and police assistants (with minimum educational requirements serving as a barrier to mobility between the latter and the former) and the use of the strike weapon. Representing different constituencies, it seems likely that the two police unions could place contradictory pressures on central government in the future. The implications of any government decision to TRANSFORMATION 33 (1997) ARTICLE WOOD challenge either union are uncertain. It is evident that industrial relations within the police service are in a state of flux, and that much uncertainty surrounds relations with senior administration, and, indeed, the bargaining process. This lack of institutionalisation makes for uncertainty, and, perhaps, a greater willingness to challenge central government in the future. Trade Unionism in the Hospitals The survey revealed that NEHAWU was the largest union operating in the health sector in the region, representing 37.1 per cent of respondents. This is followed by SANA, with 22.9 per cent. Although, in the past, more of a professional organisation than a union, it has begun the process of establishing a shopfloor organisation. Interviews with SANA office-bearers revealed that that association has begun the process of establishing a shop floor organisation with conventional shop steward structures. The first shop stewards are undergoing training in basic industrial relations/organisational skills. However, its attempts to hold talks with other unions over possible co-operation have been rebuffed, revealing something of the differing interpretations of the relative professional status of nurses versus other workers. SANA is followed by the PSL with 8.6 per cent and Hospersa with 7.1 per cent. Almost one quarter of health workers are unorganised, reflecting somewhat higher degree of union penetration in the policing sector. In addition to the above unions, interviews with union officials revealed that there are at least another eight unions either representing workers at hospitals not covered by the sample, and/or who have begun the recruiting members. However, it is evident that these 'other unions' have yet to make significant inroads in the region. Table 3 compares union membership with occupational status. T«W«3 Occup. N/R Un- Semi Skilled Super Clerical Other Totals Status Skillad Skilled visory Union Not union 0 0 0 64.7 11.8 17.6 5.9 24.3 member NEHAWU 0 19.2 23.1 34.6 3.8 7.7 11.5 37.1 Hosperea 0 0 0 0 40 60 0 7.1 PSL 6.7 0 16.7 0 16.7 50 0 8.6 SANA 0 0 0 81.3 18.3 0 0 22.9 Totals 1.4 7.1 10 47.1 12.9 15.7 5.7 100 10 TRANSFORMATION 33 (1997) WOOD ARTICLE It is evident that the different unions represent rather different categories of members. NEHAWU seems to have a following amongst all categories of employees, but is strongest amongst the unskilled, semi-skilled and skilled (the latter includes nurses). Indeed, all the unskilled workers interviewed proved to be NEHAWU members. NEHAWU faces particularly intense competition in the skilled category with SANA, and has yet to make serious inroads into the ranks of clerical and supervisory workers. It is unclear whether NEHAWU could successfully market itself in the more senior occupational categories and retain its following amongst the unskilled. In contrast, Hospersa seems confined to supervisory and clerical workers. It evidently has a rather limited following, and it seems unlikely that it has the capacity to organise the major strike it has threatened in the region. The PSL has a following amongst clerical, supervisory and semi-skilled workers. The latter would include certain categories of technical staff. Interviews with PSL official revealed that union intends to expand in all categories of employment, and it claims that membership has recently expanded rapidly. However, given the divisions in NEHAWU between nurses and unskilled workers, it remains unclear if a union predominantly representing those employed in clerical work could also effectively voice the concerns of the unskilled. Meanwhile, as expected, SANA'S membership is confined to the skilled and higher categories of employment. It should be noted that, as was the case with the police, there was a strong relationship between ethnic origin and occupational status12. Eighty per cent of respondents who were English-speaking, and 77.7 per cent of Afrikaners, were in supervisory or higher occupational categories, in contrast to only 19.2 per cent of Xhosa-speaking respondents. Thus, it is evident that the different unions not only represent different skill categories, but tend to draw on members from different ethnic backgrounds. There is little doubt that NEHAWU could erode the power bases of the other unions if affirmative action led to the increased upward mobility of its members into more senior occupational categories. However, as is the case with the police, there are certain minimum educational requirements for different occupational levels. Consequently, the main beneficiaries of affirmative action may be Africans recruited from outside the service, rather than those presently occupying menial positions. However, in either case, the unions organising workers in the more privileged occupational categories will have to provide tangible benefits to their members if they are to survive changes in the ethnic composition of the workforce. Interviews with union officials revealed that many workers in the sector, in fact, hold overlapping union memberships, and they will approach the unions alternatively, depending on their particular concerns or requirements. It seems TRANSFORMATION 33 (1997) ARTICLE WOOD that many workers believe that some of the unions are more effective in certain areas than others, and adopt something of a 'cafeteria' approach in choosing avenues to voice their demands. The research revealed that despite different constituencies, the health unions all had, (or in the case of SANA were in the process of establishing) shopfloor organisational structures: Table 4 Respondents have N/R Yes No Totals shopstewards at their workplace Union Membership Not union member 5.9 88.2 5.9 24.3 NEHAWU 0 100 0 37.1 Hospersa 20 60 20 7.1 PSL 0 100 0 8.6 SANA 31.3 18.8 50 22.9 Totals 10 75.7 14.3 100 The differences in the SANA figures (Table 4) reflect that it is still in the process of establishing shopfloor structures. There was also considerable diversity in the responses of Hospersa members. Again, that union is still in the process of building shopfloor organisation, and there seems still some uncertainty amongst membership as to what this would entail. There is little doubt that NEHAWU has the best established shop floor organisations. However, in-depth interviews revealed several disputes between shop stewards, officials and the rank-and-file. It is evident that channels of accountability are not necessarily as functional as might initially seem to be the case. Interestingly, whilst 70 per cent of health workers believed that they should have a right to dismiss a shop steward that did not carry out their wishes, only 5.7 per cent reported that a shop steward had actually been dismissed by workers at their workplace. This figure is far lower than is the case in the private sector (Maree and Wood 1994:12-20). This could be partially a reflection of the relative recency of unionisation. One half of all health workers had joined a union within the last four years. However, it does seem that the health sector unions wil! have to devote much of their energies to developing shopfloor organisation. Indeed, only 20 per cent of respondents reported that they had elected their shopstewards by means of secret ballot. Most commonly (50 per cent) this was 12 TRANSFORMATION 33 (1997) WOOD ARTICLE done by show of hands, whilst many respondents were uncertain as to how shopstewards were actually elected. Limitations in shopfioor organisation may make it difficult for unions to mobilise shopfioor support for national campaigns. Conversely, it may be more feasible for workers to resort to wildcat action on the shopfioor than gaining the backing of national officials. Such outbreaks of industrial conflict are much more difficult to resolve, and could highlight divisions in the labour movement, communities, and, hence, the tripartite alliance. The mode of respondents (47.8 per cent) believed that shop stewards should consult workers every time they acted on workers behalf. Conversely, only 4.3 per cent thought that this was not necessary. It is evident that workers expect shop stewards to accurately carry out their wishes, even if there are not totally effective mechanisms for recall. The majority (61.4 per cent) of respondents reported that they attended union meetings once a month. Only the ununionised never attended union meetings. In other words, it is evident that all the health unions have a functioning internal means for reporting back. Nonetheless, only a minority (28.6 per cent) of respondents had participated in its formulation, whilst one half thought that their union sat on any community or sectorial forum. A large majority of respondents (78.6 per cent) believed that workers would always need trade unions to protect their interests. Only 5.7 per cent felt that this was not the case, the remainder being undecided. In other words, most felt that their need for trade unions was not a temporary one, a phenomenon of the apartheid and transition eras. It is evident that public sector unions have indeed become a feature of South African life. Just 7.1 per cent of respondents felt that it would be possible for the government to make decisions without consulting them. Indeed, most respondents seemed intent on holding the government to many of its promises, and they indicated that they would not hesitate to resort to strikes or ongoing mass action should the government fail to deliver. Whilst an earlier survey revealed similar intentions amongst workers in the private sector (see Maree and Wood 1994), the latter are constrained by the need to improve competitiveness to ensure their long term employment. In contrast, workers occupying the lower occupational categories in the state sector have historically had lower wages, and have legitimate expectations that the state will seek to compensate for its discriminatory employment practices of the past. The situation is particularly pronounced in the former homelands. These regions have disproportionately large bureaucracies in relation to present needs. Whilst nominal wages were often lower, promotion was considerably more rapid, with lower tax rates and generous allowances. Already, the threat of introducing the service conditions operative in greater South Africa has sparked a wave of TRANSFORMATION 33 (1997) 13 W00D unofficial strikes involving health workers in the Transkei. Should the expectations held by those occupying positions in greater South Africa not be fulfilled, it is extremely likely that similar instances of industrial conflict could take place in greater South Africa. A larger proportion of respondents (81.4 per cent) believed that trade unions should be represented at meetings of senior administration than was the case with the police. This probably is a reflection of the higher degree of unionisation than in the police. Whilst union representation may allow for a more co-operative relationship with management, and, hence, central government in the future, interviews with union officials revealed that many existing union-management forums have had mixed results. In part, this was due to divisions in the health unions themselves. Seventy-one per cent of respondents supported unity between the various health unions. Only 8.6 per cent were opposed to union unity, the remainder being undecided. However, as is the case with the police service, certain factors count against the emergence of a combined health union in the future. Most notably, there are the divisions around occupational categories. On the one hand, there are those who occupy the lowest ranks, were, in many cases, denied upward mobility on racial lines, and who would expect some redress in the future. On the other, there are those who occupy a relatively privileged position, who would be anxious to safeguard their status in the future. However, given such overwhelming support for union unity, a combined health union cannot be ruled out in the future. Such a union would place immense pressures on the government - it would be forced to offer greater career prospects to those in lower ranks, yet be unable to cut back at more senior levels, the only solution being further expansion of the bureaucracy with a concomitant (and unsustainable) increase in government spending. Respondents were roughly evenly split over whether central administration sees unions as opposition (40 per cent believed this to be the case; 34.3 per cent not). In other words, workers are divided as to whether the government as both employer and representative is an ally or opponent. It is evident that the situation is in a state of flux. Such divisions bode ill for unity, and, again, raise the possibility of unofficial strike action. The mode of respondents (61.4 per cent) believed that public sector workers should have the right to strike. Indeed, several strikes involving health workers have already taken place in the Eastern Cape. All have either been unofficial, or under the leadership of NEHAWU. There is little doubt that the state would be extremely vulnerable should a region-wide strike break out in the hospitals. This will severely constrain the state's capacity to curtail spending. In contrast, whilst 14 TRANSFORMATION 33 (1867) WOOD ARTICLE several of the other health unions have threatened to resort to strike action, none have done so. On the one hand, this may be a reflection of the organisational weakness of these unions, and the relative vulnerability of their membership. However, collective action by these unions cannot be ruled out in the future, and they continue to actively pursue their interests through bargaining at a range of levels. This again will greatly constrain the state's capacity to restructure the health sector. Conclusion In contrast to the private sector, where unionisation is comparatively well developed, public sector unionism is still in its formative stage. Indeed, the growth of public sector unionism is inherently bound-up with South Africa's broader democratisation. However, despite this growth, these sector unions remain somewhat fragmented. Many divisions along occupational lines exist as borne out by the survey of health workers. There remains a partially racial division of labour within the public sector. Those occupying the lower occupational categories will invariably seek redress for past discrimination and improved material conditions, whilst those in slightly better positions will seek to preserve their status, placing contradictory pressures on the state. This is reflected in divisions within the public sector unions, most notably in the health sector, boding ill for future union unity in this area. Vigorous public sector unionism is a relatively new phenomenon in South Africa. Collective bargaining arrangements have yet to become institutionalised, and a wave of unionisation is invariably followed by an upsurge in industrial conflict. Again, the surveys would seem to indicate that shopfloor organisation is not as well developed as in the private sector. Should this be the case at the national level, this could well increase the possibility of increased unofficial or semi-official collective action. In many respects, the emergence of vigorous unionism in the public sector is the invariable outcome of the upsurge of independent unionism in the private sector in the 1970s and 1980s. On the one hand, this represents a further step in the direction of a more inclusive and representative industrial relations system. In most Western countries, there are well developed public sector unions that have assumed an increasingly important role with the growth of white collar vis-a-vis blue collar unionism. The extent to which public sector unionism has taken root in a historically marginalised province such as the Eastern Cape would seem to indicate that South Africa is not immune from a global trend whereby the public sector becomes the 'key source of future innovation and radicalism' within the union movement (Hince and Harbridge 1994:240). TRANSFORMATION 33 (1997) 15 ARTICLE WOOD REFERENCES Baskin, J (1991) Striking Back a history ofCOSATU, Johannesburg:Ravan. Du Toit, MA (1976) South African Trade Unions, Isando:McGraw-Hill. Govenunent Gazette 16861 (1995), Pretoria:Government Printer. Hince, K and R Hartiridge (1994) 'The Employment Contracts Act:an interim assessment', New Zealand Journal of Industrial Relations, 19,3:235-253. Keet, D (1992) 'Nurses:are they a problem or a bridge to uniting health workers'. South African Labour Bulletin, 16,7:50-55. Maree, J and G Wood (1994) 'Union Democracy', Unpublished paper presented at the Annual Conference of the South African Sociological Association, University of Natal, Pietermaritzburg. Midgley, R and G Wood (1996) "The South African Police Service in Transition:attitudes, perceptions and values of police personnel in an Eastern Cape community', in L Nieuwmeijer and J Olivier (eds) Conflict and Peace Research:South African Realities and Challenges, Pretoria:HSRC Publishers. Mtshelwane, Z (1993) 'Recognizing POPCRU', South African Labour Bulletin, 17,6:70-77. SALB see SALB Editors. SALB Editors (1992) 'The Public Sector', South African Labour Bulletin, 16,7:48-9. SALDRU (1991) Directory of South African Trade Unions, Cape Town. Newspapers Argus (Cape Town). Cape Times (Cape Town). Daily Dispatch (East London). Mail & Guardian (Johannesburg). Plainsman (Cape Town). Sunday Independent (Johannesburg). NOTES 1. The financial support of the Albert Einstein Foundation is gratefully acknowledged. 2. The PSA has also, in recent years, adopted a more outspoken stance. 3. A list was obtained of hospitals. After a number were randomly selected, they were contacted by the researchers, and access to conduct the research was requested. In most cases, this was granted. In a minority of cases, however, this was denied, and it is appreciated that this could have impacted on the rcpresentivity of the research in that hospitals, more conservative managements could have been excluded. In addition, divisions within the major independent union in the health sector resulted in a number of workers being unwilling to be interviewed. Once access had been gained, trained interviewers randomly selected respondents. 4. This component of the survey was conducted in collaboration with Professor Rob Midgley, under the auspices of the Police Research Unit, Rhodes University. A detailed account of the general attitudes and perceptions of police members in the District is contained in Midgley and Wood (1996). 16 TRANSFORMATION 33 (1997) WOOD ARTICLE 5. Here, the interviews were conducted on a station basis. This survey was conducted through probability sampling, using the area sampling method. Area sampling involves systematic sampling, where each unit constitutes a collection of elements. The advantage of systematic sampling is that the sampling frame is drawn up during, rather than prior to the interviewing process. The overall population was divided up into a series of identifiable strata, defined in terms of different police units, in order to encompass police officers from all the major divisions of the service present in the region. These constituted the primary sampling unit. Thereafter, interviewers selected respondents systematically (see Midgley and Wood 1996). 6. The final sample size was 260. 7. During the surveys, the interviewers explained the reasons for, and objectives, of the research, introduced themselves, provided details as to the research team and the supporting institutions, answered any questions, assured respondents that their individual identities would not be released, and informed them that the findings of the research would be made available to respondents on request. In most cases, attitudes to the research were extremely positive, although, as noted above, some health workers refused to be interviewed as a result of internal union disputes. After the survey was conducted, the information was processed on IBM-compatible personal computers, using the Statgraphics(R) programme. 8. These allegations surfaced in the Sunday Independent in October 1996. 9. This term is used in the statistical sense to denote the response category that occurs most often. 10. A Kruskal-Wallace analysis revealed a significant difference between racial/ethnic origin and rank level. The test statistic was 24.48048, with a significance level of 5.50707E-5. 11. The former Prisons Services (now Correctional Services) had an unenviable reputation for racist employment practices. 12. Kruskal-Wallace test statistic 15.9413, significance level 7.012977E-3. TRANSFORMATION 33 (1997) 17