Many commentators, including Deputy President Kgalema Motlanthe, blame last year's Marikana massacre on the migrant labour system. They claim that the system is intact after 18 years of democracy.
But a study of the recruitment system used on the Rustenburg mines gives a different picture.
South African mines have always relied on large numbers of unskilled labourers and the Rustenburg platinum belt is no exception. Sustaining a large workforce while ensuring handsome profits has meant mines have kept wages low.
Recruitment systems became central to supplying cheap labour. In 1901 the mines created Wenela, later Teba, agencies that recruited in Mozambique, Lesotho, Swaziland and the South African homelands. Teba delivered thousands of workers whose contracts ensured their return to rural homesteads.
In the 1980s and 1990s the National Union of Mineworkers (NUM) recruited large numbers of migrants, including in Rustenburg. It opposed the migrant labour system and had considerable success in improving wages and job security, but less in dismantling the hostel system.
Despite these changes, it was only in 2005 that Teba broke with the Chamber of Mines to become a commercial operation. It continues to recruit mainly permanent labour for mines and, for many, its offices are still the gateway to employment.
Yet recruitment patterns have undergone massive change. About 15 years ago 60% of South African mine-workers were foreign and 40% South African. Today 30% are foreign and 70% South African, with half being rural and half urban. In short, Teba recruits mainly South African workers, both migrants and locals.
Two factors have been important in this shift.
First, the Mining Charter under the new Mineral and Petroleum Resources Development Act calls for employment promotion and the advancement of the economic welfare of mining communities. The Act also requires companies to submit "social labour plans" to obtain mining licences; these must include information on the impact of mining on communities, including numbers of jobs created.
Second, the new Immigration Act encourages a reduction in foreign labour and its replacement with local workers. Mines must obtain special exemptions and pay a levy for every foreigner employed.
In the mid-1990s Mozambicans were the largest group of miners, totalling 80 000. Now only 30 000 remain and their numbers are dropping. Basotho and Swazis have also declined steeply. Teba, however, retains its migrant-sourcing monopoly, recruiting about 240 000 workers in South Africa, 35% from the Eastern Cape.
Its recruitment role has also declined. In the late 1980s the gold industry started its decline with 200 000 retrenchments. Today, such experienced workers often bypass Teba and seek work directly on the platinum mines.
Some pay bribes for work. An AngloPlat worker described paying R500 for a certificate stating that he was qualified for a job. "This happens often and the mines know about it, so they train you in case you get injured. It is not like before when Teba recruited and trained."
The reduction in migrants means that sourcing labour is now only 10% of Teba's activities. Hence it has diversified to offer a range of migrant services. It also faces competition from the new recruiters: labour contractors, or brokers, who provide short-term labour. By the mid-1990s, brokers had made significant inroads in Rustenburg and today about a third of workers – about 67 000 – are contracted.
The post-apartheid government has tried to transform the migrant labour system through such laws as the Mineral and Petroleum Resources Development Act, the Labour Relations Act and the Mine Health and Safety Act. The labour-broking system, however, has brought to the surface some of the contradictions of these well-intentioned laws.
Mines now have to accommodate the Labour Relations Act, but employers see the Act as giving legitimacy to brokers, who become the primary employer. Many retrenched gold miners and low-skilled young workers register with contractors.
Young workers, in particular, are recruited by the unregistered "bakkie brigade", who pay as little as R60 a day and may demand a R150 registration fee. Workers recruited in this way enter very short-term work, making union recruitment impossible. Larger contractors offer a total service to the mine, from sourcing to supervision and payroll management. But, unlike Teba, they mainly recruit locally.
Mines use brokers to circumvent unions, as one broker explained: "Mines don't want to go through all employment procedures, especially when they want to subcontract sections of a mine for a short time."
Another broker said: "We have no union members and don't bargain [for] wages, because our service-level agreements with companies have their own rules and pay is stipulated in the contract."
Some workers on longer-term projects join unions, but mainly for protection against unfair dismissal. This means that brokers can keep wages low, offer competitive rates to mines and make their own profit.
Permanent labour employed directly by mines is extensively unionised, so it appears that mines are adhering to labour laws, whereas in fact a large part of their workforce is weakly organised or unorganised, brokered workers.
Through brokers, management has fragmented the labour force. When permanent rock drillers revolted against the NUM and demanded higher wages last year, brokered labour was not included in their demands.
"Permanents" fear that, if unions negotiate for all mine workers, this will lower their wages and so they lack the resolve to fight for brokered labour. Wide pay differences result.
In 2012, a permanent AmPlats rock driller earned an R8 804 monthly basic, plus a R1 000 allowance, whereas a permanent Implats rock driller earned R6 540. A brokered driller, on the other hand, earned a R3 060 basic. Contracted underground sanitation assistants earn a pitiful R1 853. In 2012 Lonmin chief executive Ian Farmer and Amplats's Neville Nicolau earned an annual R17.8-million each.
Mines benefit even more from brokers' low unionisation levels, because they can discard workers by not renewing the service agreement with contractors. Recently union negotiations reduced threatened job cuts at Amplats from 14 000 to 3 000 – a gain partly achieved by terminating contracted labour.
In the agreement with mines, brokers undertake to observe safety laws because inspectors can close unsafe mines. Thus brokers hide injuries and instruct workers not to claim compensation in case the contract is cancelled. Under the Mineral and Petroleum Resources Development Act, mines must ensure contractors abide by the law, but this makes mines both referee and player, and compromises their oversight.
Contract workers also excavate dangerous areas not mined by permanents, often with inadequate equipment, as a permanent commented: "We often lend them chain block lifts to raise heavy rock, which they lift with their hands."
Workers under pressure
So what does this say about the dismantling of the migrant labour system?
Mines no longer need a centralised service. But they still benefit from low wages created by the migrancy system, which lay behind last year's ferocious pay strikes.
Labour's new mobility, coupled with high unemployment, has also benefited employers; at the same time, the use of more local labour is cutting its cost.
Mines also shortcut government's localisation push when it suits them. They hire experienced foreign labour through the back door when migrants pose as locals to get employed by brokers. An Implats manager says: "Contractors use many more migrants."
Implats defines "local" as people living within 60km of the workplace. But companies often refer to workers from North West province as "local" even if they come from Taung – 410km away!
For workers, the fracturing of the migrant system has had contradictory spin-offs. For South Africans, the repeal of apartheid laws has brought open labour-market competition free of fear of arrest and deportation to ethnic homelands. Reduced foreign migration has also eliminated much job competition.
But the focus on local recruitment threatens jobs for workers from elsewhere. Parts of the Transkei have been devastated and the economies of labour-sending countries have suffered.
Labour broking also reproduces the low-wage migrant system, with which it merges. The overt coercion of apartheid mine capital has morphed into a more hidden form, which still puts workers under enormous pressure.
Kally Forrest is a research associate and the Ruth First fellow of the Society, Work and Development Institute. This is an edited version of her Ruth First Memorial Lecture.