Article Labour Market Policy, Flexibility, and the Future of Labour Relations: the case of KwaZulu-Natal clothing industry Caroline Skinner and Imraan Valodia Introduction The issue of labour market policy generally, and more specifically the effect of South Africa's progressive labour market legislation, has been an important aspect of the debate on South African growth policy. A number of commentators (see, for example, Fallon and Lucas 1998, Lewis 2001) have argued that unrealistically progressive labour market policies, and the resultant rigidities in the labour market, are a critical barrier to employment growth. Others (see, for example, Nattrass 2001) have argued that the broader growth strategy, that of promoting high productivity, is fundamentally flawed in a labour surplus economy such as South Africa. A commonly cited response to these arguments is that of the 1996 ILO review of labour market policy in South Africa (Standing et al 1996). This report, based on the South African Enterprise Labour Flexibility Survey conducted in 1995, suggests that there are, in fact, high levels of flexibility in the South African labour market. An important area of policy that has not been analysed is firm-level responses to labour legislation - how, at the level of the firm, enterprises restructure and reconfigure their production processes in order to cope with supposedly onerous labour legislation. This article attempts to address this gap in the literature through an analysis of enterprises' reconfiguration of labour usage in the clothing industry. The research reported in this article is located within a recently established three-year research project based at the School of Development Studies, University of Natal, which aims to explore formal-informal economy labour market dynamics. Thus, mis article represents an initial and impressionistic foray into a complex set of TRANSFORMATION 50 (2002) ISSN 0258-7696 66 The case of KwaZulu-Natal clothing industry ideas and relationships that will be investigated further over the coming years. This article is based on interviews with key informants - union officials, employer representatives, bargaining council investigators, local government officials and other organisations concerned with those working in the clothing industry. Our research to date has been based largely in the Durban and surrounding areas of KwaZulu-Natal (KZN). Hence, the analysis that follows makes reference to developments in Durban and KZN province. Our analysis suggests that, although firms are not fundamentally changing their production technique, they are indeed making far-reaching changes to the manner in which labour is used in the production process. Firms are increasingly reconfiguring their use of labour by subcontracting production. We argue, however, that firm responses are largely pernicious - aimed at by-passing and undermining labour legislation - without in any manner really changing relationships at the workplace. We suggest that these developments have the potential seriously to undermine the collective bargaining system in South Africa. Much of the research that is critical of South Africa's progressive labour legislation, for example that of Lucas and Fallon (1998) and Lewis (2001), assumes that firms abide by labour legislation and collective bargaining agreements, and that where this is not the case the authorities are able to enforce the legislative provisions and collective bargaining agreements. It is also often assumed that South African trade unions are powerful enough to represent and promote their members' interests and are thus able to enforce collective bargaining and legislative provisions even when the authorities are unable to do so. For example Fallon and Lucas (1998), in seeking to explain the high levels of unemployment in South Africa, find that trade unions and bargaining council agreements in South Africa explain about a quarter of unemployment. Our initial findings indicate that assumptions of this sort are highly questionable - employers seem to be able to bypass the legislation and collective bargaining agreements with relative ease. This poses particularly difficult challenges for policymakers, for the authorities, and for the trade union movement. The article also raises some questions about the articulation of different components of South Africa's growth strategy — we examine some of the contradictions between trade and industrial policy and labour market policy. The new labour relations system introduced in the mid-1990s drew on the spirit of an era of negotiated settlements, and sought to foster the 57 kinner and Imraan Valodia t of voice regulation. Voice regulation seeks to promote the ement of labour relations, and restructuring processes through ing between conflicting interests. Proponents of voice regulation ise that labour and workplace change embodies conflicts of interests, are best managed through legitimisation and institutionalisation (see ng 1999). Tripartite institutions like the National Economic pment and Labour Council (NEDLAC) and legislation such as the r Relations Act (which encourages centralised bargaining at an ry level and the setting up of work place forums) were designed to te voice regulation. Our research demonstrates that, in the case of the ng industry, trade liberalisation though not as extensive as is popularly d, combined with weak institutions and low levels of legislative ement, is undermining the concept of voice regulation, since important sts are opting out of institutional arrangements. view of developments in the KwaZulu-Natal clothing industry ing manufacturing in KZN can be traced back to the early 1920s. Up the 1960s there were large firms that were the sole suppliers to the chains. Given tariff barriers there was little incentive for local ers to import. In the early 1960s the retailers started to source their fabric and designs and increasingly were looking for factories who prepared to cut, make-up and trim (CMT) their fabric to their patterns. mber of smaller manufacturers thus came into being. This dual ture of big and small firms remains in the industry today. According e Natal Clothing Manufacturers Association (NCMA, 2000:1) CMT panies form the majority of the individual clothing companies, although do not employ the majority of clothing workers. the 1980s decentralisation incentives were introduced. This led to a ber of the larger Durban manufacturers relocating part or all of their ufacturing activities to decentralised areas - Ezakeni, Ezikaweni, marsdale, Isithebe and Newcastle. Once the incentives were ontinued many firms moved back to the urban centres. he clothing industry in the Greater Durban area reached its peak in 0 when there were approximately 450 firms employing between 45 000 49 000 people. At this point, tariffs were on average 90 per cent on hing and it is estimated that local retailers were sourcing 93 per cent of r goods from the domestic industry (NCMA 2000). From the early 0s onwards, however, government embarked on a policy of re- agement with the global economy and a reduction of tariffs. 58 The case of KwaZulu-NataJ clothing industry South Africa's offer to the World Trade Organisation provided for an eight-year tariff phase down period for the clothing industry. The clothing industry was one of two industries (the other being motor vehicles) that were granted a longer rationalisation programme that differed from the five-year period that applied to other sections of the manufacturing industry. South Africa's tariff phase down commitment to the WTO for clothing is shown in Table 1 below. At the aggregate level, the actual tariff phase down for clothing has, as Table 1 shows, been slower that the WTO commitment. Table I: WTO Clothing Tariff Phase Down Year 1995 1996 1997 1998 1999 2000 2001 2002 WTO Tariff 74% 68% 54% 50% 46% 42% 37% 33% Actual Tariff 84% 78% 72% 66% 60% 54% 47% 40% (Source: IDC, quoted in Cassim and Onyango 2001:2 and CLOFED 2000:88) Although South Africa has significantly liberalised the trade regime in line with the WTO commitment, albeit with a lag of about two years, the system of protection remains very complex, and protection levels are still relatively high, particularly in clothing and textiles. Van Seventer (2001) provides a comprehensive analysis of tariff rates and the tariff phase down. He highlights the fact that liberalisation has slowed in the last couple of years. Further, the clothing industry remains a protected sector with relatively high tariff levels and a significant number of tariff peaks and tariff lines. Using measures of effective protection, he estimates effective protection in clothing to be somewhere between 98.8 per cent and 50.7 per cent,1 depending on the method of calculation used.2 Tariff levels portray an incomplete picture of the levels of import competition. A key concern for clothing firms is the high level of illegal imports that have entered the local market because of lax customs controls (exacerbated in part by the complexity of the tariff structure). Furthermore, the import-export complementation scheme that operates in the industry has the effect of increasing the levels of import competition for those firms that have not been able successfully to diversify into export markets. Notwithstanding the relatively high levels of protection that still exist, clothing firms have struggled to adjust to higher levels of import competition. In 2001, the retailers arc estimated to have imported between 45 per cent and 47 per cent of their goods. In other words only between 55 per cent and 53 per cent of goods are being sourced from local manufacturers. The need 59 e Skinner and Imraan Valodia mpete with imported garments has generated a massive downward ure on price. Although export volumes have increased, this has not ened fast enough to arrest huge firm closures and job losses. With the a Growth and Opportunities Act (AGOA) that grants South African ing, textiles and footwear manufacturers duty free access to markets United States, exporting is likely to be a growth area. These exporting rtunities have in the main not as yet been realised. ZN clothing manufacturers, who tend to focus on cheaper market ents, have been particularly hard hit. CMTs have been confronted the most price competition. Downward pressure on price has a cularly negative impact on labour in industries like clothing given that a high proportion of total input costs are labour costs. It is estimated 30 per cent of input costs in clothing manufacturing go to the neration of employees. This is in contrast to a more capital-intensive r like textiles in which an estimated 18 per cent of total input costs are ur costs (CLOFED 2000:91). he recent period has been characterised by significant levels of nchments. The Southern African Clothing and Textile Workers Union CTWU), which represents most workers employed formally in clothing s, has collected data on retrenchments. S ACTWU (2001:87) calculates in a two year period from July 1,1999 there have been 22 756 jobs lost e clothing industry throughout South Africa. Their latest Congress ort (2001:42) indicates that most jobs have been lost in the KwaZulu- l region. Of all 32 SACTWU branches, their Durban Central Branch rienced the highest number of enterprise closures.3 They however note there have also been significant retrenchments in decentralised areas,4 es in decentralised areas are substantially lower, with clothing workers etimes being paid less than R70 a week. This indicates the extent to h the industry is under pressure. verall figures on employment in the clothing industry, however, do not ssarily reflect this level of job-shedding. Table 2 below, for example, ests that job losses were most significant in the period 1996-97. reafter, job losses have probably been ameliorated by some level of job tion in the informal economy. 60 The case of KwaZulu-Natal clothing industry Table 2 : Number of Employees in the Clothing Sector, Sooth Africa 1993-2000 Year Total EhjJojnieiit 1993 124295 1994 124538 1995 134945 1996 149 219 (see. i) 1997 136824 1998 133 699 1999 131491 2000 138349 (Source: House and Williams quoting Stats SA figures) Note 0 The increase in 1996 is the result of inclusion of employment numbers in the TBVC states. Whilst there may be some debate about total employment in the sector, as table 3 shows, there has quite clearly been a dramatic informalisation of empoyment.5 More than half of those employed in the clothing industry are in informal jobs. Responses to trade liberalisation - the rise of formal / informal economy dynamics Key informants made the point that increased import competition has led to a fragmentation of the clothing industry in KwaZulu-Natal, and that this a ^ i n T TSt 3 C T *? t h C D u r b a n a r e a - M a Q y f a c t o r i e s hav<= "located again to decentralised areas. Some manufacturers have relocated to Even w T " ^ °°TleS e s p e c i a " y M a l a w i > Lesotho and Mozambique. Even when firms have relocated they often maintain certain operations in Durban fil,a' ? " m t e r v i e w e e c i t e d a n «™npto of a previously large ^ T Z L T DOW m a n u f a c t u ™ S *» four locations - central ^ i ^ ' a P r e v i o u s l y decentralised area north of S t i / l m i m a n a 8 e m e n t *•» had a range of options with P 8 CTS ? - ° rthan , an lssue derS *•* require Skillcd pricearepIacedintheir workers andcentral Durban where Sf ^ b e producedataeiowestpossiwe ssr 61 Caroline Skinner and Imraan Valodls Table 3: Formal and Informal Employment, South African Clothing, 2000 Occupation Sector Formal Informal Total Fibre prepares Count 515 515 % within 100 100 % within 0.9 0.4 Weavers, knitters & Count 6344 10701 17045 related %wijhin 37.2 62.8 100 % within 11.2 15.5 13.6 Tailors, dressmakers & hat Count 22348 42575 64923 makers % within 34.4 65.6 100 % within 39.5 61.7 51.7 Textile, leather &related Count 7712 379 8091 pattern % within 95.3 4.7 100 % within 13.6 0.5 6.4 Sewers, embroiderers & Count 12431 11938 24369 related % within 51 49 100 % within 22 17.3 19.4 Upholsterers & related Count 7227 3456 10683 % within 67.6 32.4 100 % within 128 5 8.5 Total Count 56577 69049 125626 % within 45 55 100 % within 100 100 100 (Source: own calculations from Labour Force Survey, Feb 2000) Those who have remained in Durban have adopted a number of strategies. There appears to be a proliferation of home-based working, particularly in the former Indian areas of Chatsworth, Phoenix and Verulam where many of those who used to be employed in the formal clothing firms live. A bargaining council investigator (Interview Sep 19, 2001) whose work, since 1995, has concentrated on identifying unregistered clothing factories explained: When factories are liquidated often what happens is supervisors and managers buy the machines and employ former factory staff to 62 The case of KwaZulu-Natal clothing industry manufacture in their garages. The supervisors will source CMT from the same people they were getting it from before. The workers will therefore often be doing exactly the same work they were doing before just under different conditions. ty officials also noted that in the last few years there has been a sudden ase in clothing manufacturing in the inner city. A Department of th official (Interview Sep 20, 2001) noted that there were buildings ghout the city that had been renovated to accommodate small ufacturing units. The plans for these alterations are often submitted as ge facilities for informal traders. When officials visited these buildings discovered that people were living and working in these units. The activity is clothing manufacture. The extent to which this type of ufacturing is linked into the formal economy still needs to be stigated. e final strategy is that of non-compliance with the Bargaining Council ements. There are a large number of firms who have continued to ate as before but have opted out of the Bargaining Council. Table 4 w records membership of the Natal Clothing Manufacturers Association MA), the employers' representative in the Bargaining Council for the ing Industry (BCCI).6 It is clear from the table that the number of oyees working under bargaining council arrangements has significantly ased. Table 4: NCMA members and their Employees, 1990-2001 Year No. of NCMA members No. of Bnployees 1990 450 45 000-49 000 1995 225 26 500 2001 65 12000 (Source: NCMA 2001:6) e Executive Director of the NCMA (Interview Sept 12,2001) estimated here is in excess of 300 employers employing approximately 20 000 e that are not complying with some or all of the collective agreements.7 the sections of the article that follow we explore, in some detail, the anisms by which firms are bypassing collective bargaining agreements iscuss the effects of this. We argue that these developments are usly undermining the collective bargaining system and leading to a increase in informalisation at the expense of formal sector employment. 63 Caroline Skinner and Imraan Valodia stations that facUitate ^ « " * ' ™* « *« Constitutional Court. dispute resolution officials can b i i * hearings in terms of their CAESAR certificates.8 What makes COFESA different from other labour consultancies is that they assist companies to restmctureAeir workforces to change employees to contractors and to outsource production to them. Section 213 (f) of the LabourRelations Act(LRA) defines an'employee' as any person, excluding an independent contractor, who works for another person or for the State and who receives, or is entitled to receive, any remuneration (emphasis added) Therefore if companies restructure their workforce into a system of independent contractors, none of the provisions of the LRA and other labour legislation apply. GOFES A firms no longer have to adhere to collective agreements on the minimum wage or contribute to pension, medical aid, sick pay, holiday pay unemployment and training schemes or funds. They do not deduct union subscriptions, supply guarantees to the BCCI nor pay overtime rates for Saturday, Sunday or public holiday work. It is estimated that COFESA firms can achieve an immediate 30 per cent reduction in labour cost. In recent years COFESA has experienced dramatic growth. A head office representative (Interview Jul 3,2001) said that in the three years he had been working for COFESA they had an average annual growth in membership of 30 per cent. Their annual report (2001:1) states that they represent over 100 000 employers' which makes them the biggest employers' The case of KwaZulu-Natal clothing industry organisation in South Africa. The founder and current director (Interview Oct 22, 2001) estimated that their interventions directly or indirectly had resulted in the establishment of 1.5 million independent contractors. He noted that it was difficult to keep exact records as often memberships lapse once members had successfully instituted the system of independent contracting. Further, there have been many reports of firms using COFESA documentation to institute this system of independent contracting while not officially becoming COFESA members. COFESA members are involved in many different sectors. COFESA's director (Interview Oct 22, 2001) noted that the key areas were clothing, leather, furniture, road freight and the metal industry. In interviews with other COFESA staff, the following sectors were also mentioned: food, farming, construction and engineering. When asked if there were certain industries that were better suited to the system of independent contracting than others, COFESA's director stated: The system can work in any industry or sector...we can work out something, we can always work out something. ; COFESA has a network of offices throughout the country. They have 25 offices, with 36 consultants listed in the 2001-2 Annual Report. They are active in the big centres and also in small towns - from Namaqualand to Tzaneen. Different offices appear to specialise in different activities. A core focus of COFESA's work in KwaZulu-Natal is the clothing industry. A head office representative said that 95 per cent of the clothing industry in KwaZulu-Natal were their members. The COFESA consultant (Interview Sep 24, 2001) working with clothing firms in the Durban offices said that he alone had over 200 firms on his books. COFESA justifies promoting the independent contractor system on the basis of support for micro-enterprise development. The director (quoted in the Sunday Tribune, Mar 12, 2000) advocates for this system on the basis that it is 'establishing micro-enterprises in a protected environment with existing markets, business skills, training and expertise'. It is also argued that independent contractors are more productive than employees and workers earn more using this system. Much of their documentation makes reference to promoting national governments macro-economic strategy, "The Growth, Employment and Redistribution Strategy', by promoting productivity growth, small business development, and black economic empowerment. 65 The case of KwaZulu-Natal dothing industry and translators to work with employees. Getting employees to understand the nature of the change has proved critical in legal cases. This was a key feature in, for example, the landmark judgement by the arbitrator SH Cnnshe in the February 2000 case between the Building Bargaining Council' and de Lange - an employer who had used the COFESA system. The Bargaining Council was arguing that dc Lange was breaching obligations arising from collective agreements. The Arbitrator awarded the judgement in favour of de Lange concluding 'there is no legal bar to a person working as an independent contractor instead of as an employee if the arrangements axe freely and voluntarily concluded' (Christie 2000:9 emphasis added). COFESA has developed a standard contract, which in the latest CODE has been translated into Zulu and Sotho. Much reference is made in COFESA documentation to the 1996 Labour Appeal Court judgement12 that states that 'the contract is the source of the relationship'. The COFESA contract is designed to set out clearly the relationship 'to avoid possible confusion with an employment relationship' (COFESA 2000:4). The first sentence of the contract clearly states that it is a contract for production and not an employment contract. It specifically states that the contractor is not entitled to protection by a trade union. The core issue from a legal perspective is how independent does someone have to be, to qualify as an independent contractor. The COFESA documentation (2000:4) quotes the Minister of Labour as identifying the hallmarks of an independent relationship as follows - there is no right of supervision, the contractor may work for another, the contractor is not required to work set or regular hours and the contractor is not paid a fixed wage but a commission or contract amount. The COFESA approach is designed to address all of these in a technical, legalistic way that minimises changes in the relationship. With respect to independence they draw on a 1996 Labour Appeal Court judgement that independence is 'relative' and that no one is totally independent. They replace the idea of supervision with quality control. The COFESA contract does not disallow contractors to work for another company but specifics that they cannot work for any company who could be a competitor. With respect to work hours COFESA states that most contractors operate during business hours and in teams and cannot come and go as they please. With respect to payment the contract provides for a 'contract amount' and payment is made on the submission of an invoice. (The CODE contains a copy of an appropriate invoice that 67 Caroline Skinner and Imraan Valodla can be used by contractors.) It is assumed that COFESA firms will continue manufacturing on the premises. COFESA however advises that contractors hire tools from the company and are charged rent. They suggest an amount of Rl a month for each (COFESA 2000:3-7). Those working under these arrangements will no longer be entitled to minimum wages, leave of any sort (public holidays, annual leave or sick leave), nor will they be able to access the benefits provided through the bargaining councils - provident fund, maternity benefits etc. According to the COFESA contract they are paid strictly for what they produce. Since contractors do not contribute to the Unemployment Insurance Fund or the Workmen Compensation Funds of the Department of Labour, they do not qualify to claim. If a dispute arises the worker will have to refer it to arbitration in terms of the contract or to the Small Claims Court or civil court. The Bargaining Council for the Clothing Industry (Natal) has paid out in excess of R1.5 million in prosecuting over 500 legal actions between March 1999 andMarch 2001. The majority of these cases have been against COFESA companies. These legal processes have resulted in Conciliations, Arbitrations, Labour Court Orders and Orders for Contempt of the Labour Court. In a typical case the Bargaining Council will give the non-compliant employer a notice for conciliation. In most cases the employer does not attend the conciliation or takes it on leave to appeal. The matter is then set down for arbitration. The Arbitrator then seeks an order that the employer register with the Council and the Sheriff then has to serve the award. The award will grant the employer more time to comply. Only once the Bargaining Council has gone through this whole process and the employer defaults can the Council make an application to the Labour Court to have the award made into a Court order. Each step in this process takes time. The conciliation and arbitration process takes between two and four months. Because of backlogs in the Labour Court, the Council will wait a minimum of six months for their application to be heard. It is only at this point that there is any legal sanction against the employer. In the meantime the employer continues operating as before. As one Bargaining Council agent (Interview Oct 26, 2001) pointed out before decriminalisation of labour legislation 'within six weeks the Council had actions, now it takes at least a year if not more'. Even once there is a court order COFESA firms frequently take leave to appeal. It is estimated that firms have a three-year window before being forced to comply. The COFESA director openly The case of KwaZuiu-Natal clothing industry admits to exploiting the pace of the legislative process. As a labour lawyer (Interview Oct 25, 2001) who has been working COFES A cases pointed out COFES A usually wins on the basis of procedural rather than substantive grounds. A Bargaining Council representative (Interview Oct 26, 2001) pointed out 'in the last three years the merits of the case of independent contractors versus employees have not yet been heard'. Further, when employers finally have no legal recourse, they close. In one of the few cases that resulted in the arrest of an employer," once the employer had served his 15-day jail sentence he closed his factory and all the workers lost their jobs. This places the union - SACTWU - in a very difficult position. This has led some to argue that the problem is in fact with insolvency legislation. Implications for collective bargaining The intervention of COFESA has seriously undermined collective bargaining arrangements, particularly in the clothing industry in KwaZulu-Natal but also in other industries. For nearly 70 years there has been an industrial council regulating substantive conditions of employment in the Natal region of the clothing industry. Voluntary collective bargaining systems however depend on three factors — compliance to negotiated agreements from employers; a trade union that is in a position to organise the majority of employees; and an enforcement system that works. In the case of the Bargaining Council for the Clothing Industry (Natal) none of these factors is currently present. As Table 4 indicates the employers' organisation, the NCMA, now represents a small portion of employers, with the majority of employers not complying with Bargaining Council agreements. The Union has experienced a dramatic decrease in membership numbers. Despite minor success in organising COFESA firms in the end of 1999, SACTWU (2001:52) concedes that they have been largely unsuccessful in organising unregistered factories in this region. The employees of non-complying factories often resist fearing that they will lose their 'jobs'. Withrespectto the enforcement system, the decriminalising of labour law has reduced the BCCI's sanctioning power. The NCMA entered the 2001 wage negotiations stating: The NCMA wishes to renegotiate all the clauses in the main agreements that in our opinion inhibit flexible work practices and place us on an inequitable footing when compared to other clothing manufacturers in the region. Anything less will lead to the untimely demise of the BCCI and the NCMA. (Smart 2001:8) 60 Caroline SWnnw «nd Imraan Valodia The 2001 agreement between the NCMA and SACTWU stipulates a wage freeze and manufacturers can vary downwards as a system of wage bands per job grade replace the minimum wage schedule. Previously a grade 1 machinist earned R421 a week. If a manufacturer did not pay this they were non-compliant. The new agreement specifies a wage band of R341 to R441 for machinists, thus allowing firms to pay within this range. The argument is that the compliant manufacturers can now compete with COFESA type firms for orders since there is some flexibility in wage rates. It should be noted that clothing workers, on the R441 rate, would on average earn R1829 a month. This is above UPE's (2001:57) minimum household subsistence level for Durban for September 2001 of R1432.85. If wages are decreased to R341 a week, workers will be earning in an average monthR1477, only just above the minimum household subsistence level. If the agreement is followed through there is nothing stopping currently compliant employers from decreasing wages. When asked about this the head ofthe NCMA (Interview Sep 12,2001) said that employers could only vary downwards if there was agreement from employees through a secret ballot. There, however, is nothing stopping employers from coercion - for example threatening closure if employees do not agree or introducing new lines paying employees on these lines less and placing the more expensive workers on short time. The NCMA were only in a position to secure this agreement because the capacity of their institution and the union and therefore the bargaining council has been so undermined by COFESA. Interviewees' views on this new agreement were mixed. Some trade union officials were very critical of the agreement and argued that the union negotiators had not canvassed members sufficiently. Further, both unionists and employers expressed doubt about whether the new agreement would succeed in reinvigorating the bargaining council by attracting COFESA firms back into the collective bargaining system. The head ofthe NCMA, for example, argued that 'the harm is done... there is now a culture of non-compliance'. The COFESA director, however, claimed they do not want to destroy bargaining councils, but rather want to weaken them to the extent that they can secure seats within the bargaining council structures. He pointed out that if there were no bargaining councils there would be 'no battles to fight' and therefore employers would have no need for COFESA's services. In KZN, for example, the NCMA is negotiating with COFESA for them to be part ofthe manufacturers association representing non-parties. 70 The case of KwaZulo-Natal dotting Indusby They currently sit on the metal industry bargaining council. Addressing the legislative loophole At the heart of the legal debate is the issue of what constitutes an employment relationship. This issue has implications for informal sector debates more generally, particularly with respect to social protection for those working in the informal economy. In COFESA cases, the Bargaining Council and the Union have argued that there is no substantive change in me nature of the employment relationship. The proposed labour legislation changes are in part targeted at COFESA. style arrangements. The proposed amendments shift the burden of proof, ^and propose a series of rebuttal presumptions as to whether or not an employment contract exists. A person is now assumed to be an employee until the contrary is proven. A person is considered to be an employee if any one or more of the following factors arc present — a) the manner in which the person works is subject to the control or direction of another person; b) the person's hours of work are subject to the control or direction of another person; c) in the case of a person who works for an organisation, the person forms part of that organisation; d) the person has worked for that person for an average of at least 40 hours per month over the last three months; e) that person is economically dependent on the person for whom he or she works or provides services; f) the person is provided with his or her tools of trade or work equipment by another person; g) the person only works or supplies services to one person." If these labour amendments are passed as they stand, there is likely to be a series of cases against COFESA firms." These will lead to important precedents with respect to the issue of what makes an employee an employee. All COFESA staff that were interviewed seemed unconcerned about these changes. The head of the Durban office (Interview Jul 20, 2001) said 'We will find other ways of doing things'. Another consultant (Interview Aug 24,2001) said,'We will just call the independent contractors something else'. The director said (Interview Oct 22,2001) that COFESA would now call 'independent contractors', 'entrepreneurs' and that the organisation was focusing on development of entrepreneurs. COFESA has 71 Caroline Skinner and Imraan Valodia inserted a training and development programme for entrepreneurs in the latest copy of the CODE. It appears as if they will continue implementing this system arguing that this is a micro-enterprise initiative. A senior Durban based labour lawyer (Interview Sep 26, 2001) argued that amending the legislation is insufficient. He questioned who would be in a position to take cases against COFESA - the Bargaining Council and SACTWU are both too weak. The workers in COFESA type firms are no longer unionised. These workers are in a weak position with respect to opposing their employers. Conclusion There is a tendency internationally, and certainly in South Africa, to overstate the growth of informalisation and flexibilisation, and also to see these developments as very recent, and unconnected to historical traj ectories in the labour market. Though the developments outlined in the article suggest that new forms of informalisation and flexibilisation are beginning to emerge in South Africa, it is important to note that the labour market in South Africa has historically been characterised by high levels of flexibility and informality. A key characteristic of the apartheid system, and the racial pattern of South Africa's industrialisation, was the highly flexible system of contract and migrant labour (see, among others, Wolpe 1972, Legassick 1974, Hindson 1987). Several micro-level studies in the early 1980s have suggested extremely high levels of informal economy activity (see for example, Cross and Preston-Whyte 1983, Wellings and Sutcliffe 1984, and Webster 1984). It is also important to note that current trends in the South African labour market towards the 'hollowing out' of firms and the increasing use of non- traditional forms of labour are consistent with international trends. Standing (1999) outlines the growth of flexible forms of labour throughout the developed and developing world. In a study of five South Asian economies (Bangladesh, India, Pakistan, Sri Lanka and Nepal), Unni (2000) reports on the massive growth in informal employment and the growth of informal employment within the formal economy. Before summarising our findings three caveats are worth noting. First, our analysis in this article is based on the clothing industry, an industry that has been particularly hard hit by South Africa's integration into global production networks. The pressure on clothing firms to reconfigure their production is particularly acute. Hence, the trends outlined in the article 72 Tta C M * of KMZufe»#Mri oMNno m*N*y may not be transferable to other industries. There u some evidence, however, that many induatriea are experiencing similar treads, albeit not with the same intensity (see, for example. Kenny 2001 on retailing, and Crush et al 2001 on gold mining). Second, our analysis is confined to Durban and its immediate surrounds. Again, we suspect that these trends apply in other centres of industrial production. Third, as pointed out earlier, our research project is very much in its infancy. A striking feature of the manner in which firms in the Durban clothing industry are reconfiguring their production is the pernicious nature of their actions. The growth in independent contract type arrangements, promoted by organisations such as COFESA, are aimed primarily at bypassing aspects of the labour legislation.1* Despite claims to the contrary, there is no evidence to support the view that these are truly subcontracting arrangements, aimed at promoting the growth of small business. Our article provides evidence that the liberalisation process, by forcing some firms to restructure in line with COFESA type arrangements, is seriously undermining the objective of fostering a comprehensive collective bargaining system based on the concept of voice regulation. The pressures of trade liberalisation are forcing some enterprises to opt out of voice regulation institutions such as bargaining councils. These tensions raise important challenges for policymakers, and also for the trade union movement in South Africa. Notes I Van Seventer estimate* effective protection based on the so-called Balassa method and the Corden method. The Corden method yields lower estimates (see van Seventer 2001:25 for detail*). 2. It is often argued that the Department of Trade and Industry (DTI) has implemented trade liberalisation faster than is necessary, ie that tariff reduction has been implemented more rapidly than necessary to meet obligations to the WTO. At the aggregate level. Table 1 suggests that this is not the case. However, the tariff regime remains highly complex, particularly for items of clothing. At a disaggregated level liberalisation in some cases may have been faster that required. This is an area that the authors will clarify in the coming months. 3. The report does not disaggregate closures by industry. Their closure figures therefore include closures in textile and footwear factories. 4. They report that of the 56 factory closures that resulted in over 100 job losses 33 were located in decentralised areas. ' 73 Caroline Skinner and Imraan Valodia 5 There has been a long history of informal distribution of clothing in KwaZulu- Natal; what appears to be increasing is informal manufacturing. 6 The scope of the Natal Bargaining Council covers the magisterial districts of Durban, Pinetown, Tongaat, Stanger, Verulam, Inanda and Pietermartizburg. These are boundaries inherited from the Industrial Council days. 7 There are indications that the dynamics in the Western Cape and Gauteng are quite different. 8 Section 200 of the LRA states' Aregistered trade union orregistered employers' organisation may act in any one or more of the following capacities in any dispute to which any of its members is a party- (a) in its own interest; (b) on behalf of any of its members; (c) in the interest of any of its members' 9 It should be noted that this figure of 100 000 includes both big and small employers as well as federations that COFESA is affiliated to. Further, as an employer of a domestic worker you can become a member of COFESA. This is also likely to inflate their numbers. 10 COFESA staff have legislative expertise and in some cases appear to have worked in the previous government. Leon Wessels, the minister of labour prior to the political transition, sits on the COFESA board. 11. It is interesting to note that the Building Bargaining Council no longer exists. The reasons for this, particularly the extent to which COFESA contributed to its demise, still needs to be investigated. 12. Judge Nugent in Liberty Life Association of Africa v Niselow. 13. See BCCI (Natal) v Snap Clothing. 14. Government Gazette No 21407, July 27, 2000. 15. It should be noted that COFESA appears to be one of many labour consultancies that have promoted the system of independent contracting. References Cassim, R and D Onyango (2001) 'South Africa's trade reform and the World Trade Organistion: background and progress', Trade and Industry Monitor 19, September. CLOFED (Clothing Federation of South Africa) (2000) CLOFED Handbook 2000/1. Bedfordview: CLOFED. Crush, J, TT Tuiicki and E Jansen van Vuuren (2001) 'Undermining labour: the rise of sub-contracting in South African gold mines, Journal of Southern African Studies 27(1). . ' Confederation of Employers South Africa (2000) CODE for Fairness and Productivity, COFESA. (2001) Annual Report 2001-2002, Mimeo. 74 The cam of KwaZulu-Natal clothing industry Christie, S (2000) Arbitration Award. In Arbitration held at Worcester to resolve me dispute between the Building Bargaining Council (North and West Boland) and Chnstoffel Steyn De Lange. Case #BNW AA 0055 1999. Cross, C and E Preston-Whyte (1983) 'The informal sector: desperation versus maxmiisation strategies', Indicator South Africa 1(2). Fallon, P and R Lucas (1998) 'South Africa: labour market adjustments and inequalities', Informal discussion papers on aspects of the economy of South Atnca, Discussion Paper #12, World Bank. Gannon, S (2002) 'Quasi informalisation in the Durban clothing manufacturers sector: a look at casualisation temporary work and labour broking'. Masters Dissertation, School of Development Studies, University of Natal, Durban. Hmdson, D (1987) Pass Controls and the Urban African Proletariat. Johannesburg: House, K and G Williams (2000) 'The state of employment in the South African clothing sector: diverting a race to the bottom'. Paper presented at the 2000 Tips Annual Forum, www.tips.org.za. Kenny, B (2001) "There is no one who loves you when you are not working: flexibility, fragmentation and the reconfigured identities ofretail sector workers on the East Rand'. Paper presented at the SASA conference, UNISA, Pretoria, July 1-4. ' Legassick, M (1974) 'South African capital accumulation and violence', Economy and Society 3(3). Lewis J (2001) 'Policies to promote growth and employment in South Africa', informal discussion papers on aspects of the economy of South Africa, Discussion Paper #16, World Bank. Natal Clothing Manufacturers Association (2000) 'The Clothing Manufacturing Industry m KwaZulu-Natal'. Mimeo. Nattrass, N (2001) 'High Productivity Now: a critical review of South Africa's Growth Strategy', Transformation 45. SACTWU (2001) 'Congress Report'. Standing, G (1999) Global Labour Flexibility. London: Macmillan. Standing G J Sender and J Weeks (1996) Retructuring the Labour Market: The South African Challenge. Geneva: ILO. University of Port Elizabeth (2001) 'The Household Subsistence Level in the Major Urban Centres of the Republic of South Africa'. Fact Paper 109: University pf Port Elizabeth. Unni, J (2001) 'Gender and informality in the labour market in South Asia', Economic and Political Weekly, June 30. 75