Home > Features > The state of labour 20 years after democracy

South Africa

The state of labour 20 years after democracy

Tuesday 27 January 2015, by Niall Reddy

Save this article in PDF Version imprimable de cet article Version imprimable

“It was one of those great historic moments,” Ruth First wrote after 60 000 black mineworkers struck for a wage increase of 10 shillings in 1946, “that in flash of illumination educates a nation, reveals what has been hidden, destroys lies and illusions.” The strike was brutally repressed by the state, leaving 9 dead and the workers’ demands were largely unmet. But its political effects were huge.

Sixty-six years later, South Africa’s mines have again become a window into the heart of our society. On August 16 2012, police massacred 34 mineworkers in front of the world’s TV cameras. Their crime had been to demand a living wage, of R12 500 per month.

When the ANC government came to power it built a vision of a free and just society that would work to erase the injustices of the past. The cornerstone of this vision was the guarantee of decent work for all.

It was moments like the 1946 miners’ strike and the many struggles that preceded and followed it that helped South Africans to understand the cold economic logic that underlay Apartheid – the system of racial labour control. The ANC’s vision relied on redeeming labour from this brutal history, and making it part of a democratic economy.

This was, lest we forget, a formidable task. South Africa’s economy bore the curse of its incredible mineral wealth. The huge reserves of gold, diamonds, coal, platinum and other precious resources gave rise to a small club of giant conglomerates that had a powerful hold over the state and influenced the way the entire country developed. The result was an economy dominated by these and closely related industries, which was highly capital intensive, and left 30% of the working population without jobs.

By the end of the 80s, Apartheid had expended its usefulness to this “Minerals-Energy Complex”, and now more than anything wanted to take part in the opportunities offered by globalization and the rapid expansion of finance. Fixing the gaping problems in the South African economy, investing in a diverse manufacturing sector and in the infrastructure to provide basic services to millions of South Africans was not on the agenda for capital.

Therefore, as many pointed out, the only road to the ANC?s vision was a direct confrontation with capital. Only by breaking the back of the Minerals Energy (-Finance) Complex, by directing the wealth of the economy to the needs of the people through nationalization and public finance, could a decent economy with decent work be achieved.

For the ANC government, Marikana was a watershed. What it revealed, in horrific fashion, was the failure of this historic task.

As the commission of inquiry proceeds, despite political interference and police intimidation, some things cannot be hidden. The brutal police response appeared as at least partly a result of political pressure. Though the faces may have changed, the Minerals-Energy Complex remains.

And though it may rely more on the threat of starvation in underdeveloped rural areas than official state policy, so does the system of cheap migrant labour from which its profits derive. Despite the promises of decent work, Marikana showed how far the current government is willing to go to defend it.

At the time of writing, mid 2014, tens of thousands of workers in the platinum sector have picked up the spear of their fallen comrades, in what is reputed to be the longest strike in South African history. This marks the most important battle in the labour rebellion that was sparked by Marikana, which quickly ignited labour unrest on the farms in the Western Cape before spreading to other sectors of the economy.

The upsurge in union militancy has also lead to a re-alignment of political affiliation, with NUMSA leading the charge to take up the radical demands of workers and forming a workers political movement. It is the first time that a prominent member of COSATU broke with the ANC and SACP, who were seen as having a close relationship with MEC capital.

Capital’s response has been to go on the offensive calling on the ANC to use the post-Marikana developments to smash unions and abandon the National Democratic Revolution in favour of the National Development Plan. This effectively completes the neoliberal transformation begun in 1995.

The purveyors of the new propaganda are claiming victory. It would be understandable if it were just business that bought into the new mantra but many academics and state bureaucrats, whose job it has been to expound the promise of formal employment, have claimed that formal sector workers are a privileged group in the new South Africa. The most common variant of this argument says that powerful unions and strict labour regulation have led to the creation of a “labour aristocracy” whose high wages prevent the creation of more jobs.

If this is the case, the current labour rebellion is impossible to understand. We are instead left with the conjured image of venal union bosses and old fashioned greed.

It is against this backdrop that we ask how workers have fared in 20 years of democracy. As the new government made the promise of a just society dependent on the creation of decent work for all, to pose this question is really to assess the success of the democratic project.

This paper uses empirical data, mostly drawn from Post-Apartheid Labour Market Survey (PALMS) to examine the conditions of workers from a variety of angles, including Employment, Wages, Job Quality, Institutions, and Unions.

Contrary to right wing commentary, we find that the revolt of workers against the existing labour market regime has a very rational basis in declining wages, rising unemployment and weakening employment security. These outcomes, in turn, are a consequence of the ANC’s embrace of the Minerals-Energy-Finance Complex economy, and the failure of workers to mount a challenge to this.

To change these circumstances workers will have to not only rebuild militant, democratically controlled unions, but to mount a serious challenge for state power.


We start our analysis by looking at employment trends in the economy. Graph 1 shows the change in employment in different sectors of the economy over the period 1994 to 2012. The main trend is a shift in employment towards the ‘tertiary sector’ – retail and service jobs. Manufacturing employment grew only slowly and moderately and the primary sector – mining and agriculture – declined.

Graph 1 Employment by sector 1994 - 2012

The biggest growth sector in terms of jobs over this period was trade which went from almost 1.65 million jobs in 1994 to just over 3.1 million in 2012, growth of almost 90%. This growth comes mostly from the retails sector.

‘’ Financial Services also grew rapidly by more than 220%. This includes people who work in banks and other financial institutions, but also those classified as ‘other business services’ which are mostly labour brokered workers – especially security guards and cleaners. Over 66% of the total employment growth in the finance sector in this period is due to labour brokered workers. (This is illustrated in Graph 2).

Graph 2 Employment in finance and labour broking ’94 – ‘12

Source: PALMS

Services was another high growth area. In 1994 there were 2.3 million service jobs, and by 2012 there were 3.04 million. All of this growth happened after 2000 ,as data show that in the late 90s service employment actually declined. Health and education workers make up the biggest part of the service sector.

Domestic services also grew hugely, by more than 380%, although some of this may be because the statistics were not as good at capturing domestic work in the 90s. In the second half of this period growth was slow but by 2012 there were over 1.2 million people employed in domestic work.

Two other growth sectors have been construction and transport, which grew by 130% and 60% respectively.

Manufacturing employment grew very slowly - just over 10% between 1994 and 2012. This means an increase of only about 160 000 jobs.

However, because of outsourcing, the statistics may underestimate employment growth in manufacturing and other sectors and overstate services. Many of the security guards, cleaners and other workers listed under “finance” in fact work in the manufacturing sector and would previously have been employed directly.

Nevertheless, job growth in manufacturing has been hugely disappointing.

Mining employment decreased rapidly in the late 90s, losing almost 200 000 jobs between 1994 and 2003 as the gold mining industry collapsed under low prices and declining reserves. Thereafter some jobs were recovered mainly due to the expansion of the platinum industry, mining employment was 523 000 in 2012.

These statistics bear witness to the dramatic collapse in agricultural employment as farms mechanized and restructured towards export markets – in 2012 there were just under 700 000 employed in agriculture, under half the 2003 level.

Many economists believe that the manufacturing sector is crucial to growth. It is through the manufacturing sector that the economy improves its productivity and engenders new technologies. Most developed economies went through a period of strong manufacturing growth before shifting to services. When developing strategies for reforming the economy around the transition time, many progressive economists pinned their hopes on a growing, diversifying manufacturing sector to generate growth and jobs. This hasn’t happened.

Instead we have a shift to the service sector dominated by retail, transport and outsourced work. In Marxist terms the retail sector does not produce value – it only facilitates its circulation. Many of the jobs in this sector are lower paid and more insecure than other parts of the economy.

More research needs to be done to understand the exact effects of these sectoral shifts on workers. Table 1 suggests that many of the high growth sectors paid lower wages, whilst the sectors which contracted as a proportion of total employment were associated with better paid jobs.

Table 1 Sectoral shifts and wage 1997 – 2011

Source: PALMS

Both the average and the median wage in the trade sector, which grew substantially and made up almost a quarter of all jobs in 2011, was around 20% below the economy wide level.

The finance sector has high mean and median wages due to rich bankers and their employees. But labour brokered work where much of the growth was concentrated has markedly lower mean (-26.3%) and median (-8.47%) wages.

Similarly, in construction, which made up 2.35% more of total employment by 2011 relative to 1997, mean and median wages are lower.

In transport, wages are generally higher – but the sector grew largely in line with overall employment growth meaning it accounted for a constant share.

An important exception to this rule is the service sector. The sector made up a large share of overall employment, 21,83% in 2011 (up by 3,05% from 1997). Both mean and median wages in the service sector are substantially higher than the whole economy level.

In manufacturing, mean and median wage levels are higher than the average for the economy. But because job growth was slower than the average, the sector made up an increasingly smaller proportion of overall employment.

Other sectors that contracted as a proportion of employment, such as mining and utilities, also had relatively higher wages.

On the other hand, the agricultural sector lost a number of jobs. These were mainly poorly paid.

More rigorous statistical work is needed to understand the effects of these shifts. But the patterns described above strongly suggests that the growing weight of retail and services and stagnation had a negative effect on wages.

The service sector generally has lower “value-added”. Retail workers are less organized than their counterparts in manufacturing and mining.

Another important trend in employment over the last two decades was the change in the composition of private and public employment. Graph 3 shows employment in these categories as an index starting in 2000.

Graph 3 Public and private sector employment ’00 – ’12 (INDEX = 2000)

Source: PALMS, author’s calculations

In the first half of the 2000s, public employment shrunk and then grew slightly so that it was basically the same in 2007 as it was at the start of the decade. Private employment on the other hand grew fairly substantially over this period by almost 18%. It grew a further 8% over the next year before collapsing after the global crisis in 2008 and stagnated thereafter, leaving employment in the private sector slightly below the level it was in 2006.

Some of the slack was picked up by the state. Public sector employment grew by almost 15% between 2007 and 2012. In 2012, 2.4 million workers, just over 17% of the total, were employed by the public sector.

The post-Apartheid labour market also witnessed a shift away from lower skilled towards white-collar work, as shown in Table 2.

Table 2 Skill composition of employment 1994 – 2012

Source: PALMS, author’s calculations

The number of white-collar jobs increased by over 90% between 1994 and 2012. The fastest growth was in the skilled category, rising by 113,3%, which includes clerks and various technical professionals. The highly skilled category, which includes a high number of managers, senior administrators and more professionals grew by 74,57%.

Blue-collar work grew much slower, at 38,8% overall. Low skilled employment, elementary and domestic occupations, was the slowest growing category at 33.25%.

As such, blue-collar work made up 5.5% less of total employment by 2012 compared to 1994. Still almost 3 out of 4 jobs in the economy are blue collar.

The main point that we observe here is that the number of new jobs have not matched that of new entrants to the jobs market. Not surprisingly then, this is reflected in a rise in the rate of unemployment (Graph 4).

Although it might seem a pretty simple matter there is considerable disagreement amongst labour analysts about how to define and understand unemployment. At a basic level, the unemployment rate is the number of people in the ‘workforce’ who do not have a job (‘unemployed’), divided by the number of the entire workforce.

One of the main disagreements is about whether to include in the number of unemployed those workers who have given up looking for work. Official statistics usually report the official rate of unemployment as not including these workers - the ‘narrow’ rate - which is always lower [1]

However, , labour research has repeatedly shown that a large percentage of people who have stopped looking for work have done so because they no longer have hope of finding a job, not because they do not want to work [2].

Anecdotal reports also indicate that many of the unemployed who have given up the search for jobs continue their search at home with phone calls or through personal networks.

For this reason, the broad measure of unemployment, which includes ‘discouraged’ workers, is the preferred rate of actual unemployment because it captures all unemployed persons who are able and willing to work.

Graph 4 shows total employment, the ‘actively searching’ unemployed and discouraged workers. It shows that total employment in the economy grew from 9.09 million jobs to 13.89 million between 1994 and 2002. The fastest periods of growth were between 1998 and 2001 (about 2.7 million jobs) and 2004 and 2008 (about 2.4 million).

Graph 4 Employment and unemployment 1994 - 2012

Source: PALMS, SARB, author’s own calculations

Between the start of the global economic crisis in late 2007 and 2010 the economy lost over 900 000 jobs. By 2012 they hadn’t been regained. This is one indication of the “extreme flexibility” of the South African labour market which allowed capital to pass off the cost of its crisis to workers.

However, between 1995 and 2003 the number of people actively looking for work and the number of discouraged workers grew faster than the number of jobs created, meaning that the broad unemployment rate increased from 29.5 to 41,5%. The narrow unemployment rate in 2003 was 30.5% - considerably lower.

Over the next five years the economy added jobs fairly rapidly, meaning that broad unemployment declined slightly to 37% in 2007.

In 2008 the Labour Force Survey changed to the Quarterly Labour Force Survey and StatsSA changed the way job seeking questions were asked. A number of people who were previously counted as discouraged were now regarded as “inactive”. This explains why the number of discouraged workers drops suddenly from 356 000 to just over 110 000. This is known as a “structural break”. It means that broad unemployment before and after the break are not strictly comparable.

After the change, which coincided with the crisis, unemployment again began to rise rapidly. At first the economy was shedding jobs before regaining them slowly. However, while the number of “actively seeking” remained constant, the number of discouraged workers, using the new, stricter definition, rose rapidly. By 2012, broad unemployment stood at 33.2% according to official statistics.

In human terms, that means 6.9 million people who are eligible to work cannot find a job in this economy. This is a crisis on a grand scale. In comparative terms, at the height of the ‘Great Depression’ in the early 1930s, one of capitalism’s worst ever crisis which toppled governments around the world, the unemployment rate in the US was 25%.

Before we even get to discuss the quality of the jobs that were actually created, these statistics show that the ANC government failed in its promise to expand the boundaries of the so-called “first economy” to encompass all South Africans by creating enough jobs.

Because they make up the majority of new entrants to the labour market each year and are less educated and experienced, the unemployment crisis in South Africa, as in most countries, is concentrated among the youth. The unemployment rate among the 16 to 34 year old age group in 2012 was 45.8%.

Women and people living in rural areas are also disproportionately affected.

The latest QLFS for 2013 estimates that 65% of the narrow unemployed are long-term unemployed – having been out of the job market for a year or more.

Right wing commentators use this situation of chronic unemployment to claim that having a job is a privilege. Other commentators, including progressives, speak of “two separate classes” within the proletariat – formal workers and the mass of unemployed- and under-employed.

According to the business-friendly media it is the high wages of the first group that is keeping the unemployed from getting work. In a previous paper we dismissed this argument and showed that higher wages are likely to lead to higher employment (Reddy, 2013).

In reality there are not two separate classes – workers suffer directly and indirectly from the employment crisis regardless of the “class” to which they are affiliated.

Marx called the unemployed the “reserve army of labour”. He described this as the chaos of capitalist development, where workers are continuously cast into the reserve army as new technologies make them expendable, pressuring them to re-skill to be re-absorbed only to be expelled at some later stage again. As the reserve army grows it depresses wages by forcing workers to compete harder for available work – offering to sell their labour for less and less.

Some have claimed that this doesn’t happen in South Africa because the labour market is “inflexible” and because business faces skill shortages and is unwilling to hire the long-term unemployed. We challenge this assertion and show here that in fact the labour market is extremely flexible.

One important study showed that wages are negatively affected by local unemployment – in the same way as in other labour markets [3]. The high degree of churning which we discuss later is also evidence of intense competition between active workers and those forced into the reserve army by capital.

What effect then has the unemployment crisis had on wages? In the following section, we take up this discussion.



Employers and their experts frequently claim that wages are too high. In 2012 Mike Schussler’s influential annual employment report entitled “South Africa can’t Afford South Africans”, sought to persuade the public to see the economy as a business where wages are a “cost”.

Schussler claimed that rising wages were making South African businesses uncompetitive and called for increases in wage inequality.

In fact real wages for most workers have hardly risen since the end of Apartheid. We show here that only wages for the wealthiest workers – mainly managers and some professionals – have increased which have been driving up “average” wages.

Graph 5 shows that average real wages across the economy were largely stagnant from 1997 to 2002 before rising. The mean wage for the whole economy grew by 37,7% between 2003 and 2011 to R6564 (2011 prices) [5]. The trend for formal sector, non-self-employed was very similar, rising from R5415 per month in 2003 to R7316 per month in 2011. Wage increases for public sector workers began slightly earlier in 2001, rising by 22% by 2011 to R8838 per month.

Graph 5 Real mean monthly wages 1997 – 2011 (2011 Prices)

Source: PALMS, author’s calculations

Average wages are calculated by adding the total wages paid and dividing it by the number of workers. Although averages are commonly represented as the wage that ‘most’ workers get, the average actually says very little about distribution. If inequality is very high, as it is in South Africa, the average can be far from what ‘most’ workers earn. Moreover, if only the wages of the richest workers increase, this could lead to an increase in the average wage even though the wages of most workers have not risen at all. Data shows that this is exactly what happened.

We can show this by dividing workers into groups based on the wages they earn. The 10th percentile (p10) shows the wage of the lowest paid workers who make up 10% of the total. The 50th percentile is also known as the median wage. It divides workers into two halves. One half earns above that wage and the other half earns below it. The 90th percentile shows the minimum wage for the top 10% of wage earners.

Graph 6 Real monthly wages by percentile 1997 – 2011 (2011 Prices)

Source: PALMS, author’s calculations

Graph 6 shows real monthly earning trends for all workers by percentile. The graph shows that not even the minimal promises of a free market economic system were met over the last 15 years – workers at the bottom half of the earnings scale received only a fraction of total wages .

Table 3: Real wage growth rates, all workers

Source: PALMS, author’s calculations

Workers towards the middle of the distribution saw the largest declines in real wages in percentage terms. The median wage in the economy decreased from 1997 to a low of R2451 in 2003. Between 2003 and 2011 it recovered somewhat to R3029, still 7,2% below it’s level in 1997. The 30th percentile wage also dropped over the 15 years by just over R200 to R1724 per month in 2011.

The poorest workers saw a slight convergence with those immediately above them. The 10th percentile wage rose from R654 to R851 per month, a 30% increase though still firmly in the region of poverty wages.

These figures show that the average increases described above were entirely driven by increases for higher earners – specifically those right at the top of the distribution. The 70th percentile wage also declined notably in the 6 years after 1997 from R5457 to R4597, thereafter recovering fairly strongly to R6084 per month in 2011. By far and away the largest gains, however, were for the top 10 percent of workers, whose wages increased by 37.7% between 1997 and 2011 to R15028.

By dividing the percentiles to get a ratio we can get a rough and ready measure of inequality at different parts of the distribution (Graph 7).

Graph 7 Real wage percentile ratios 1997 - 2011

Source: PALMS

The p50/p30 ratio was relatively constant suggesting that wages in the middle grew at relatively the same rate. The ratio of the 50th to the 10th percentile decreased showing that low wages caught up somewhat with wages in the middle.

The ratio of the 90th percentile wage to the 70th increased slightly and then stagnated showing slightly growing inequality at the top of the distribution. The biggest divergence was the top from the middle, with p90/p50 going from 3.3 in 1997 to 4.9 in 2011.

We can also show what happened by looking at the change in real wages by skill level (Graph 8). This time we use the mean as there is not too much inequality within each skill category. Similar patterns to previous graphs are discernible.

Graph 8 Real monthly mean wages by skill 1997 – 2011 (2011 Prices)

Source: PALMS

Lower skill gradients experienced stagnant real wages.Elementary occupations and domestic workers (low-skilled) received more or less the same mean wage in 2011 as in 1997, around R2200. Semi-skilled workers saw wage declines at the beginning of the period before a slow increase to R4924 per month in 2011.

Skilled workers saw moderate gains from R7647 to R9126 over the 15-year period. These were realized between the period 2001 to 2007, after which wages declined.

However, again it was only highly skilled workers at the top of the income distribution that had significant earnings increases. Wages for this groups, mostly managers and other professionals, increased by 71.3% to R18661, almost twice that of the next skill category.

These findings are corroborated by various firm level studies. In their comparative studies of the automotive industry, Black and Hasson note that “the most striking feature about the labour market in South Africa is not so much that wages of production workers are higher than competitors (although in many cases they are), but the exorbitant costs of managers and skilled staff.” [6].

A major 2007 report for the World Bank found that the median monthly wage for a manager in South Africa – about $1850 - was over twice that of Poland and three times that of Brazil. Moreover, SA managers were found to earn nine times as much as unskilled workers, whereas in Brazil and Poland it was only three times as much. Similar patterns were found for other skilled and professional workers. Their study concurred with our findings that ”high wages in South Africa appear to be mainly due to high wages for managers and professionals and not to high wages at the bottom of the income distribution.” [7]

Thus data from StatsSA completely contradicts the arguments of business that workers have been big winners in the post-Apartheid economy. If there is a “labour aristocracy” it is made of ‘highly skilled’ professionals and managers. It is the wages of this group that show significant gains. Workers in most of the other categories saw almost no improvement in real wages in the decade and a half after 1997.

When it began in the late 70s most countries to which it spread experienced declining labour shares of GDP. This refers to the share of all the value produced in the economy in a year that goes to workers in the form of wages and other remuneration.

Graph 9 shows the labour share of GDP for South Africa between 1994 and 2011. Here we have netted out the wages of managers and supervisory workers who are in actual fact more like bosses.

Graph 9 Adjusted labour share of GDP 1994 - 2007

* Adjusted for self-employment
Source: Onaran and Galanis, Is Aggregate Demand Wage-Led or Profit-Led? National., PALMS and author’s calculations

It stands as a strong indictment of South Africa’s record after 20 years of democracy that workers have received a smaller and smaller share of the wealth that is produced each year.

In 1994 workers got 68% of the value produced, by 2007 it was 59.4%. The wage share of the actual working class shrunk even more dramatically by almost 11% from 56% to 49.3%. That means for every R100 produced in 2007, workers got R11 less than they would have had had the share been the same as in 1994.

The much hoped for redistribution that should have succeeded the end of Apartheid did not materialise .

Of course the above analysis misses the most popular narrative of modern inequality which is the staggering wealth accumulated by the richest of the rich, the “1%”.

Our data focuses only on wages and excludes other sources of income like social grants, which are important for the poor, and capital ownership, which is very important for the super wealthy.

The French economist Thomas Piketty has written a book which many consider to be the defining text on inequality in the modern capitalist era. In Capital in the 21st Century, he shows that capital has a permanent tendency to reproduce inequality as the return on capital, r, exceeds the rate of growth in the economy, g.

Piketty’s book has some theoretical problems but data he assembles are very important. He avoids the problems with household surveys (like we have used here), which are not good at capturing the extremes of the distribution – the richest and the poorest – but uses tax data instead.

His data on the US contains a rather startling observation, one which somewhat escapes his grand narrative – that 2/3rd of the increase in income inequality over the last decades is attributable to drastic wage increases for the highest earners – chiefly the new class so-called “super-managers”.

Various studies have found that South Africa has amongst the highest CEO remuneration levels in the world. This is an issue that has come under increasing scrutiny as strikes in the mining sector catapult wages and the ethics of remuneration to the centre of the public debate. One study has noted a three fold increase in executive salaries in the country’s 50 highest earning companies, from an average of R15m in 2005 to R49m in 2014 [8].

More work needs to be done on the relative importance of capital and wages in understanding South African super-incomes.

Piketty’s data suggests a substantial rise in the top 1% share of income in South Africa from 9.85% in 1990 to 16.68% in 2011, putting it on par with some of the world’s most unequal nations like the US. However, James K Galbraith has criticized excessive faith in tax data, noting that they are only as good as the collection systems that underlie them and that Piketty’s results jar with other studies that have found South African and Brazilian inequality, for example, to radically outstrip that of the US [9].

In studies conducted by Dick Forslund he found huge disparities between high net worth individuals found in the databases of the financial sector and high income earners registered with SARS. This casts some doubt that Piketty’s data captures the full scope of top earnings in SA [10].

Elodie Morival, a student of Piketty, compared the tax data with national income surveys and found that they told the same story, though this may reflect the limitations of both rather than their accuracy – a clear picture of inequality in South Africa may have to wait for greater stringency from SARS for high-end dodgers [11].

Job quality

During Apartheid workers had few rights and almost no legal protection. The workplace was a crucial site for creating a pliant and subjugated population, from whom the maximum value could be extracted for the lowest pay.

When the ANC assumed power in an Alliance with COSATU one of the chief ways it proposed to create a new society was the promotion of ‘decent work’.

This is a term that was popularized by the International Labour Organization in the 1990s. It embodies the notion of work that is secure, formally regulated and offers benefits with upward mobility for the worker. This is usually based on the so-called “Standard Employment Relation” (SER) which refers to a direct, contractual relation for permanent employment between the employer and the worker.

One of the main features of neoliberal globalization that emerged out of the capitalist crises in the 70s and had a major impact on the world in the next four decades was changes it wrought in the world of work. In short, this involved a shift away from the SER towards news forms of flexible, less regulated and indirect “a-typical” employment relations.

Neoliberals have celebrated this – claiming that the removal of stability unleashes creative forces by forcing all of us to sell ourselves on the market like commodities. This allowed businesses to be more “flexible” in a rapidly changing world.

Of course labour has been critical of the effects that precarious employment has on the job conditions and lives of workers.

One influential author, Guy Standing, who worked for the ILO for many years, coined the term the “precariat”. He claims this is a new emerging “dangerous” class that has no interest in investment in social stability and is prone to subversive politics.

Much of the literature on precariousness has a strong Northern bias. There is a tendency to emphasize the “newness” of these changes or their departure from a supposed ‘norm’ of employment – typically the type of Fordist standard employment relation which encompassed mostly white male workers in rich countries during capitalism’s Golden Years.

But for the South, precariousness has always been the norm. At the time of the famous strikes in Durban in 1973, it was reported that half of the 200 000 workers involved were ‘contract workers’.

Protected SERs never really existed during Apartheid except for a portion of white workers – a small sliver of the proletariat.

Through the struggles which revived the union movement in 70s and 80s workers began to win greater legal recognition. It is at this time that we begin to see the clear emergence of labour broking and other means of evading regulation.

The freedom to organize and the greater protection and rights that workers won through successive amendments to labour market institutions and law is one of the great prizes of the transition to a formal democracy.

However, these amendments were inspired by an idea of the ‘standard employment relation’, which was out of touch with global trends in employment. They did not present serious barriers to capitalists seeking to avoid stricter regulation by switching to various forms of “a-typical” employment.

For example, the Labour Relations Amendment Act sanctioned labour broking by allowing a third party to remain as the legal employer.

The massive shift to a-typical employment by capitalists in the last few decades meant that precariousness did not end with Apartheid. Unions have been ineffective at contesting this. Precariousness and a-typical employment is usually described using considerable jargon which seldom has a precise meaning.

In broad terms, casualization refers to changes that affect the duration of work, generally by moving from permanent, full-time positions to ones that are part-time or based on fixed period contract.

Externalization refers to processes that involve a third party in the employment relation. This could be through outsourcing where a service or part of the business is contracted to a different company or through a labour broker .

Informalization refers to changes that make work less regulated by labour law or market institutions.

South Africa has been subjected to all of the above on a huge scale. However, because data are inadequate and the phenomena do not have strict definitions, it is hard to give exact measurements of the scale of precariousness.

A related development in the South African economy is the huge informal sector. Similarly imprecise, this is a term that was popularized in the 70s to refer generally to areas of the economy that do not have conventional capitalist features. That is, work is not capital intensive and the means of production are owned by those who work them. Often it refers to non wage-earning or self-employment. In South Africa small traders, craft production and construction services make up a large part of the informal economy .

The informal sector has often comprised between 20 and 30% of total employment. In an important speech in 2003, Thabo Mbeki explained the “two economies” thesis that was to have a big impact on policy. He described South Africa as having a large, capital intensive “first world” economy based on mining and manufacturing with standard employment relations, and a separate “second economy” comprising informal and subsistence activity.

But several researchers since then have shown that the two economies are anything but separate. They are deeply linked through production, the circulation of goods and employment, the last an inseparable part as workers in the massive “reserve army” which along with the pockets of informal activity help to suppress wages and ensure reproduction of class relations in the capitalist order .

Graph 10 shows the distribution of the workforce in 2001 and 2011.

Source: PALMS, authors calculations

In this scheme, the informal sector is defined as businesses that are not formally registered or do not pay VAT.

Between 2001 and 2011, 8% more of the workforce was included in employment of some sort. However the proportion of the permanent formal sector employment remained fairly constant at about a third of all workers.

The graph suggests that the expansion in employment was firstly in formal non-permanent work, which grew from 4 to 9% of the workforce. Secondly, informal employees went from 10 to 15% of the workforce, with the total informal sector employing 22% of workers in 2011.

The trend has thus not been an expansion of a so-called “first economy”, but an increase in various forms of precarious and informal work.

As ‘informality‘ has increasingly become part of the main capitalist economy in most countries, the 17th International Convention of Labour Statisticians suggested a definition that focuses on the nature of the job rather than the employer. Using South African statistics, some researchers define informal employment as jobs which have neither a written contract nor benefits (pension and paid leave).

Table 4 Benefits and regulation as a proportion of employees

Source: PALMS, author’s calculations

Table 5 shows informal employment as a proportion of employees in the economy decreased between 2001 and 2011 by over 10% to 30.23%. Employees with a written contract, by contrast, grew from 54.9% in 2001 to 79.7% in 2011.

However, the proportion of workers receiving benefits increased at a much slower pace or declined slightly. In 2011 just 67% received paid leave – up from 59,9% a decade earlier. Only 31.9% of workers received medical aid contributions from their employers, more or less the same proportion as in 2001. In 2011 just under half of all workers got pension contributions, less than 2001 when it was 51.6%.

The slight growth in some benefits and the spread of written contracts in the second half of the decade is likely attributable to the growth in public sector employment (see graph 3).

While these statistics suggest that more workers are covered by a formal contract they do not capture changes in the nature of work contracts that may offer workers less protection or are less able to access labour advisory or litigation services

Labour force surveys show that an increasingly lower proportion of employees are engaged in permanent work (Table 5). Between 2001 and 2006 the proportion dropped by more than 5% to 75.44%. The share of employees with temporary work and those in casual or seasonal employment increased by 2.7 and 2.3% respectively.

Table 5 Work duration as a proportion of employees

Source: PALMS, author’s calculations

StatsSA changed the way it measures employment duration with the introduction of the QLFSs from 2008. In 2011, a total of 25.8% of employees reported being engaged in work that was either of a limited or unspecified duration. In the same year, just over 64,2% of employees had permanent contracts.

These statistics are an indication of the increasing casualization of the workforce. The standard employment relation, which is supposed to underpin the common conception of decent work, applies to an ever smaller share of workers.

However, the figures here do not capture the full extent of increasing precariousness or the many forms it has taken.

An important example is labour broking. Labour brokers are known formally as ‘temporary employment services’ but in many, perhaps most, cases they are used for jobs that are essentially permanent or full time.

Labour broking can best be characterized as the triangular employment relation. In practice, brokered workers usually work directly for a client company, which defines their day-to-day activity and determines their pay and conditions of work. However they are legally employed by a third party – the labour broker – who is sometimes also responsible for hiring.

Labour broking confers an advantage to the employer by making it harder for workers to access labour law services. Often the confusion generated by the triangular employment relation is an instrument of the control exercised over workers.

When a worker attempts to proceed with court action against her employer, for unfair dismissal, for example, the boss simply claims that he is not the legal employer and refers the case to the broker, who turn claims he/she is not involved. In a great many cases, workers are not even aware that they are legally employed by a third party.

Labour broking also divides the workplace – on a single factory floor there could be a dozen different legal employers with which workers are required to bargain.

This is one reason labour broking has been so popular amongst employers as a means of weakening worker organization and increasing flexibility. The industry is reportedly worth R46 billion annually.

Because of the confusion, few workers report being employed by a labour broker, although the figures in Graph 2 (a proxy for labour broking) suggest that at least 1 million workers are involved in triangular employment. Bargaining councils and broking companies themselves give similar estimates – but some experts put the figure at twice that [12].

Since it deprives workers of the protection of the labour laws of this country, COSATU has called labour broking a “modern form of slavery” and demanded it be banned. The ANC has made many promises to stop labour broking but has not moved on the issue so far. Alongside the imperative to divide and undermine worker organization, the shift to precarious employment has meant been driven by desire to conform to new modes of “flexible” and just-in-time production and to maximize short-term gains to shareholders. This has shifted the focus from increasing productivity and creating long term value to minimizing costs – chiefly labour costs.

The ANC’s liberal, pro-market macroeconomic policies have encouraged this labour regime.

As a result the ideal of stable, protected employment has not been widely achieved. For most workers in South Africa precariousness remains the reality. The celebrated protections in the labour law make little difference.

Precarious workers are much harder to organize. They are more subject to abuse by bosses and are less able to fight for higher wages and better conditions. The next section explains how unions have failed to represent precarious workers.

Like elsewhere, some analysts claim that the working class is now divided with diverging political interests – between the ‘precariat’ and the unemployed on the one hand, and the formal sector workers on the other.

A point that some analysts ignore is the issue of churning. Studies have shown that South Africa’s labour market is very unstable with workers moving frequently between various forms of employment and joblessness.

Table 6 Churning in the labour market 2002 - 2004

Source: Valodia and Devey (2012)

Valodia and Devey showed that more than half of workers that featured in the Labour Force surveys between 2002 and 2004 had changed their work status. Only a very small proportion of workers remain securely in formal permanent positions.

This is another example of extreme flexibility in the labour market, despite what employer claim.

Precariousness affects all workers. By dividing workers it weakens the entire labour movement.

Labour law and instiutions

What of the labour laws and specialized institutions that were created after 1994 to deal with labour relations? After Marikana and the massive wild cat mining strikes in 2012, the farm worker rebellion in Western Cape in December that year is has been clear that the labour laws and institutions established after apartheid are in crisis. All of these events completely bypassed the fifteen year old Bargaining Councils and to a large extent also the established trade unions.

A liberal democracy allows juridically “equal” partners to negotiate their opposing positions. Workers and employers are anything but equal in capitalist economic systems. There are, nevertheless, several pieces of legislation that were introduced in the mid to late nineties which were progressive in intent. The Employment Equity Act is one important example. The Basic Conditions of Employment Act is another.

It is frequently claimed that South Africa has an overregulated labour market that contributes to labour inflexibility. In fact most attempts to construct international comparisons have demonstrated South African labour law to be relatively liberal. South Africa’s latest score on the OECD employment protection legislation index was 1.25 (for 2008), the lowest of all G-20 emerging markets and well below the OECD average.

Benjamin, Bhorat & Cheadle (2009) found, using World Bank data, “that in most measures of labour regulation … South Africa is not an extraordinarily over-regulated (or indeed under-regulated) labour market”. Perhaps even more worringly – there is enormous evidence that labour laws and regulations, such as minimum wages, are systematically ignored and avoided by employers.

Perhaps the most galling evidence of South Africa’s extreme labour flexibility was the fact that the economy lost over 1 million jobs between the onset of the global crisis in 2008 and mid 2010.

The bargaining council system (renovated from the old Apartheid industrial councils), voluntary associations between employers and registered unions with provisions to extend to agreements to non-parties, has been a consistent target for those lamenting the lack of flexibility in South African labour markets. But BCs have weakened over the decades as part of a marked trend away from centralized bargaining. Veteran trade unionist Neil Coleman (forthcoming) argues that:

While centralised bargaining is critical for the labour movement . . . , it is not in its current voluntaristic form, able to drive the transformation demanded by the South African situation. Indeed the fragility of collective bargaining institutions can be used by employers to attack wage levels. [13]

Over 80% of applications for exemption from BC extensions are granted – suggesting that employers have little reason to be threatened by agreements that cannot be forced by their own workforce [14]. The sectoral determinations instituted where BC are lacking may have had some successes in raise wages but have hardly been sufficient to provide a real challenge to poverty wages.

A range of institutions were also born out of the idea of the “social compact” between Capital and Labour. It is useful to briefly refer to some of the most important among these.

 [15] was established in 1995 to facilitate cooperation on economic, labour, and development issues; promote economic growth, participate in decision-making and social equity and work for the resolution of socioeconomic disputes. The work of the Council is conducted in four chambers: the labour market chamber, the trade and industry chamber, the development chamber, and the public finance and monetary chamber.

The chambers report to a management committee which oversees the work programme and administrative issues. It brings together representatives from all sectors of society who debate and try to reach consensus on social and economic policy issues in what the body terms “social dialogue”.

Through the chambers, unions have a platform from which they could advance their claims on the wealth their members help to produce. The chambers allow employers and employees to present all relevant data - such as contained in this paper – in an open and transparent forum where it can be disputed or debated.

This is at least the idea. Anecdotal evidence suggests that unions, in particular, have not used this platform to their advantage. Since some years, Business as well as the Treasury has chosen not to take important issues to NEDLAC. The so called “youth wage subsidy” is one example. A general idea has spread that NEDLAC isn’t working or has lost in importance. Its usefulness or legitimacy is put in question.

What has happened in society – or not happened – since 1994 suggests that NEDLAC is in crisis because its foundational idea is in crisis. The two sides, “Business” and “Labour” that are supposed to compromise around a common cause of “what is best for the country”, are too far apart. The global crisis for capitalism crisis since 2008 has deepened the divide. In short: there are no deals brokered at NEDLAC.

The Commission for Conciliation, Mediation and Arbitration (CCMA) followed in 1995 as a dispute prevention and resolution body. It replaced the Industrial Court. It is not controlled by any political party, trade union or business organisation.

Its policy-making structure is an 11-member governing body. It has three state representatives, three representatives of organised labour, three representatives of organised business, a chairperson and the director of the CCMA.

The CCMA’s main brief is to:

• Mediate to prevent and settle industrial disputes;

• Conciliate workplace disputes;

• Arbitrate disputes that remain unresolved after conciliation; and

• Facilitate the establishment of workplace forums and statutory councils.

CCMA provides a free service, even to unorganised workers. About 25 percent of the cases deal with domestic workers. [16] Labour journalist Terry Bell notes: “Employers with deep pockets can still ‘thumb their noses’ at the CCMA by dragging out the dispute process when they have not got their way. This is done by taking matters to the labour court and further if necessary until financially exhausted worker claimants fall by the wayside.” [17]

During the 2014 platinum strike CCMA intervened without success. Anecdotal evidence given to AIDC from Western Cape after the farm worker strike in 2012 point to that the politics of who is the judge is of crucial important in every arbitration case. That politics and not the letter of the law are crucial is also obvious in the case of Department of Labour.

The case of Section 27 in the Employment Equity Act indicates that the lack of inspections on site or absent implementation of progressive labour laws – like the maximum 9 hour work day (with the aim of implementing the 8 hour work day) in the Basic Conditions of Employment Act – isn’t only a question of a too small bucket to ladle with in an ocean of labour law breaking and abuse of workers.

Section 27 of the Employment Equity Act of 1998 (“EEA”) was drafted and adopted to address income inequality and to reduce the income gap between the highest paid managers and the rest. “Where there are disproportionate income differentials, a designated employer must take measures to reduce it progressively. […] Such measures may include collective bargaining, compliance with sectoral determinations (section 51 of the Basic Conditions of Employment Act); the application of norms and benchmarks recommended by the ECC, relevant measures contained in skills development legislation, and any other appropriate steps.”

The provisions of Section 27 have never been used by either the government, the Department of Labour (DOL) or the Employment Conditions Commission (ECC) to achieve its aims. ECC has never recommended any “norms and benchmarks” for reasonable wage disparities between workers and managers to the Minister as it should do. [18] The report form that employers must fill in when reporting income inequality has no space for the average wage in different wage groups. Wage inequality between the groups isn’t measured between them. Indeed, the DOL uses the “Paterson scale” in all its reports. It divides the work place in six wage levels based on “decision making” and was first used by Anglo American in the 1960’s to make a “scientific” argument for the wage gaps between managers and workers.
For politicians (national, provincial and local), senior bureaucrats and highly paid executives in the private sector, the failure to implement section 27 reflects the protection of self-interest. It is harder to explain the inaction of unions on such an important piece of legislation. It was Cosatu that pushed Section 27 into the EEA in 1998, but all unions demand wage increases in percent “across the board”. This increases income inequality on the work place.

To decrease inequality, wage demand must be formulated in Rand across the board, as was done by Amcu during the platinum strike. This is not a new idea in South African trade unions. It has in fact been practiced before by the National Union of Mineworkers (NUM).


The union movement which emerged slowly out of the important strikes in Durban in the early 70s and the labour market reforms that followed at the end of the decade were a central part of the anti-Apartheid movement.

The Congress of South African Trade Unions was formed in 1985. In __ it took the decision to form an Alliance with the African National Congress and to give first priority for the national democratic struggle as a precursor to the fight for socialism.

When the ANC assumed power, COSATU reaffirmed this strategy, aiming to use its Alliance with the ruling party to influence policy in favour of workers.

The democratic dispensation removed all the barriers to association, allowing workers to organize freely.

In the new, more sympathetic political climate it was hoped that unions would expand to include greater numbers of workers, especially the traditionally excluded. A stronger union movement could fight for higher wages and better conditions for all workers.

The union movement has faced serious challenges since democracy. The Marikana massacre and the current divisions within COSATU have drawn attention to some of these difficulties.

Whilst the number of union members has grown, the proportion of workers who are organized (union “density” or “coverage”) has shrunk consistently. Unions are increasingly dominated by higher skilled, public sector workers and the most exploited and precarious have fallen through the cracks.

Moreover, the union movement has lost the militancy it once had. the political compromises that COSATU made, have, to a large extent, caused a growing “social distance” that separates leaders from the rank and file.

Graph 11 shows that union density varies considerably by sector. In every sector the proportion of union members decreased between 2001 and 2011.

Graph 11 Unionization by sector, all employees 2001 - 2011

Source: PALMS, SARB author’s calculations

The mining sector, where employees live and work closely together in large numbers, has historically had very high unionization rates – close to 80%. Services and utilities, which are comprised of large numbers of public sector workers such as teachers and municipal workers, also have high rates of union density.

Unions have never succeeded in penetrating the agricultural sector, where workers are less skilled, live far apart and often depend on farmers for housing. In construction, work is often temporary or piece meal and unions have had a hard time achieving stable growth.

Graph 12 Distribution of union members by sector, all employees 2011

Source: PALMS, SARB author’s calculations

In the period from 2001 and 2011 unionization rates decreased in all sectors. In most cases the decline was between 2% and 5%. In services, 5.71% fewer workers were union members in 2011 compared to 2001.

Moreover, employment growth was generally faster in sectors with lower union density. Employment in manufacturing, which makes up a large proportion of union members, grew by only 2.84%.

Two of the fastest growing sectors, construction and finance both have extremely low union density rates.

The relatively high growth of employment in services was extremely important in sustaining unionization rates across the economy. In 2011, service sector workers comprised almost half (45%), of all union members. This may also explain the growth in employment in the public sector .

Table 7 shows that whilst unionization rates in the formal private sector decreased notably in the decade leading up to 2012, the trend in the public sector was different.

Table 7 Union coverage, formal sector employees 2001 – 2012

Source: PALMS, author’s calculations

In 2001, 29.2% of private sector employees were members of a union. Over the next 6 years the number of private sector union members grew by 456 000 – slightly slower to employment growth leading to a small decline in union density. Between the start of the global crisis and 2010 private sector unions lost 303 000 members although unionization coverage stayed relatively constant – suggesting that union members were similarly affected to non-unionized workers.

Over the next two years private membership continued to decline slowly to 1.76 million members even as the economy regained jobs. This led to a drop in union density to 25.9% in 2012. Unions have not managed well during the crisis.

Union density in the public sector is much higher, and increased from 68.6% in 2011 to 74.2% in 2010 before dropping again to 71.6% in 2012. Public sector workers now made up almost half, (48.7%), of formal sector union members in South Africa, with 1.67 milllion members.

Despite declines in unionized membership, Overall union density in South Africa still compared favourably to international standards in 2010 (Graph 13). The average for OECD countries was 17.6%, considerably lower than South Africa’s 38.6%. However some evidence suggests that Latin American and other middle-income countries had far higher union coverage.

Graph 13 Union density OECD and selected countries, 2010

Source: OECD, ILO

Another notable post-Apartheid trend has been a shift towards higher skilled, white-collar workers amongst unionized workers. Table 8 shows that white-collar employees increased their share of unionized workers versus lower skill grades by 5% between 2001 and 2011 to 33.5%.

Table 8 Skill composition of unions, all employees 2001 - 2011

Source: PALMS, author’s calculations

The best organized workers in the economy were skilled workers, mostly clerks and technical professionals, who increased their union coverage over the decade to 53.4%. The union density of highly skilled workers decreased between 2001 and 2011 but was still relatively high at 29.8%.

Although they made up almost two thirds of union members, only 25.6% of blue-collar workers were unionized in 2011, a drop of 3.7% from 10 years before. This followed from decreases in union density for both semi- and low- skilled workers of around 3.3 and 2.3% respectively. In 2011 only 14.6% of low-skilled workers were members of a union.

In 2011, only 64.25% of workers in the economy reported holding permanent contracts, but within COSATU the figure was 93% (according to the 2012 survey). The growing cohort of precarious workers are falling through the cracks and not joining unions [19]..

These shifts were highly concentrated in COSATU. By 2009, the share of public sector workers in the federation was 40%, up from 9% at the start of the 1990s.

The federation claimed small successes in organizing the most vulnerable – by 2008, 10.2% of the membership were part-time or temporary workers – a 3% increase from 4 years earlier. But far more needs to be done if COSATU is to truly represent the working class.

Bosses have actively used precariousness to divide workers – convincing a “core” permanent workers that improving conditions can only be afforded if the precarious are excluded. Unionists have sometimes been complicit in this – refusing to devote resources to workers whose membership is more vulnerable or to pursue new strategies for fear of “informalizing” their membership.

Old forms of industrial unionism may need to be changed in order to adapt to restructured workforces and include a-typical forms of employment, but union leaders sometimes have a vested interest in keeping things as they are.

Economists often measure the effectiveness of unions by using statistical techniques to get a “union wage premium”. This can be thought of the as the difference in wages between workers union members and non-members that are otherwise the same in terms of individual characteristics (like age, education, race and gender) and job characteristics (like occupation, firm size and benefits).

Table 9 Union wage premia 2001 - 2011

^ Union wage premium times union coverage
*, **, *** Statistically significant at the 10%, 5% and 1% level
Source: PALMS, author’s own calculations

Table 9 shows that union wage premiums for all skill grades were increasing between 2001 and 2011. However for white-collar workers, data were not ‘statistically significant’ meaning we can’t get an accurate picture of wage effects for these groups.

Union wage premia were largest for unskilled workers although relatively constant over the decade – going from 19.4% in 2001 to 20% in 2011. For semi-skilled workers the change was more substantial, an increase of 5.4% to 10.8% in 2011.

However, these methods are quite problematic and often what is reported as the effect of unionization is due to different things that data doesn’t easily capture, like unionized workers occupying different jobs to non-union workers. Moreover, the effects of unionization cannot only be gleaned by looking at members wages alone.

To understand how unions have fared we should look at a broader range of data, that economists usually ignore.

Along with a very different social composition, the nature of unions and their function have changed dramatically in the democratic era.

COSATU’s strategy of influencing the state from within has required a different kind of leadership with specialist skills. The new and more complicated institutional environment and the emphasis on negotiation has shifted the focus of union activities from the shop floor to the boardroom. Workers are less able to be directly involved.

According to the latest COSATU survey, around 42% of workers said they had not attended a union meeting in the last year and around 35% said they had not had shopsteward elections in the last four years [20].

Although a majority of workers say they continue to directly elect shopstewards instead of having them appointed, they exercise less direct control over their immediate representatives. In 1994, 72% of respondents said that shopstewards should do only what membership tells them to do. In 2008 it was 46%. [21]

The weakening of worker control and the co-operative approach to the ruling party and business encouraged some to use the unions as a “transmission belt” to better things. In the latest survey 52% of workers said they had seen a shopsteward promoted to a managerial position [22].

At the highest level this is reflected in the personal trajectories of people like Gwede Mantashe, Kgalema Mothlante and Thulas Nxesi who have gone from the union movement to senior positions in government.

Union leaders have a strong incentive to toe the party line and to not be too confrontational towards bosses – which may conflict with the interests of the membership.

This professional leadership leads increasingly different lives to the average COSATU member. Many of them are paid very well, have stable jobs with good conditions including cars, and do not experience the same exploitation that workers do. COSATU has called this “social distancing” – without strong worker control it makes it harder for leaders to represent their membership and tends to make unions more conservative.

Another thing that Marikana revealed was the dangers of co-opting worker organizations and turning them into sweetheart unions. In the worst cases NUM had been partly absorbed into mining companies with shopstewards taking positions in human resources department and senior leaders drawing additional salaries [23]. They had ceased to properly represent workers.

The conservatism of South African unions is reflected in industrial action statistics. This may be a surprise to many.

The media regularly reports that organized labour is “holding South Africa to ransom” or “bringing the economy to a standstill” with excessive industrial action. We are led to believe that unions and striking workers are standing in the way of growth. In the World Economic Forum’s Global Competitiveness Report South Africa usually features at the bottom of 144 nations in terms of labour-capital relations based on surveys with business.

However recent research shows that South African workers are less strike prone and that strikes are shorter when they do happen, relative to other countries [24].

The researchers calculate a measure of strike intensity for a country as the number of striking workers in a year divided by the number of formal sector workers. For South Africa it was 1.1% in 2003, 7% in 2007 and 2% in 2011. On average, for the decade up to 2008 on 2.8% of formal sector South African workers went on strike.

Their data shows that this is considerably less than many other countries such as Spain (7.3%) and Argentina (30.7%) (Graph 14).

Graph 14 Strike intensity 1998-2008

Source: Bhorat, Naidoo, and Yu, “Trade Unions in an Emerging Economy.” pp13 based on ILO data

They also measure the ‘depth’ of strikes in South Africa by averaging the percentage of strikers’ workdays lost per year. This tells us how long strikes in South Africa tend to last. Between 1999 and 2008 strikers on average lost 3.77% of their workdays. Again, this is low by international standards (Table 10).

Table 10 Strike depth 1999-2008

Source: Bhorat, Naidoo, and Yu, Trade Unions in an Emerging Economy. pp14 based on ILO data

All BRICS nations for which there is data had greater strike depth than in South Africa. In Brazil strike depth was 12.3 times as high as in South Africa and in Turkey it was 5.24 times higher.

Thus relatively few workers down tools each year and when they do it lasts relatively shorter. Business’s perceptions of a militant, recalcitrant workforce probably has more to do with Apartheid era norms of labour control than anything else.


The state of labour in South Africa should be unsettling for anyone who expected the transition to political democracy to end one of the worst systems of exploitation in a modern industrialized economy. The labour practices of the Apartheid era are still in place, despite the sugar coating of a few statutes that carry the imprimatur of protection and remedial action. After two decades of political democracy, our “progress” in the field of labour can be summarised as follows:


  • An unemployment rate of 45.8% for the 16 – 34 year age group (2012); for ages 15 – 25 it is 56%
  • A shift to the low-wage retail and service sectors
  • Growth in financial and business services but predominantly in labour brokered services
  • A mismatch between the jobs created and new entrants into the labour force


  • A slight increase for the poorest paid,
  • A decline of the median wage level
  • Stagnant real wages for lower skilled workers
  • A substantial increase for the top 10% of earners
  • Moderate gains for skilled workers

Job Quality

  • A general increase in precariousness, through casualization, informalization and externalization.
  • A much smaller proportion of the employed hold permanent contracts
  • A rise in labour brokered jobs (Triangular Employment) to between 1 and 2 million


  • Centralized bargaining has been weakened and ineffective in raising wages
  • The CCMA does not effectively protect workers because formal equality before the law ignores the much greater financial and social power that employers wield
  • NEDLAC is essentially non-functioning due to deep divergences between labour and business
  • Section27 has been completely ignored by employers, politicians and labour representatives
  • By international comparisons labour regulations in SA is relatively weak


  • Large drop in private sector union coverage
  • Employment growth was faster in sectors with low union density
  • Shift towards public sector workers within unions
  • Unionization among higher skilled, white collar workers is growing
  • Very few of the precarious are members of unions

What might workers consider to be an effective program of action in the coming years?

South Africa’s economic performance since 1994 has been disappointing to many. Poverty and inequality have worsened to the point where one in four persons go hungry (that’s more than 14 million people!) and the country is considered to be the most unequal in the world. But it is the level of unemployment that distresses most South Africans. In the last instance we must understand this unemployment crisis based on the specific nature of Minerals Energy Complex capitalism in South and the way it has evolved under neoliberal globalization. Apartheid bequeathed an economy heavily concentrated in mining and related, highly capital intensive sectors to a mass of the workforce was “surplus” – a permament reserve army. Liberalization has not fundamentally altered this structure – it has allowed corporationso f the MEC to globalize and restructure, facilitated the financialization of the economy and led to growth of trade and retail – but no new, employment creating manufacturing base has been created. The growth potential of this new economy is directly dependent on its competitiveness with the rest of the world – which has been ensured by low wages rather than innovation and productivity gains.

In the last instance – it was the policies chosen by the government that allowed the Apartheid economy to evolve in this way. Policies reflect several things about a government: how it conceptualizes relations between state, market and labour; which of the great classes in society it is beholden to; how it sees the extent of its sovereignty. In essence, policies are the arms of the ideological corpus of the state. In the past 20 years, the government of South Africa has demonstrated which ideology it prefers to be guided by. It has the dishonorable distinction of having imposed a structural adjustment on its own people that went even beyond what the World Bank or the IMF demanded. It would be intellectually dishonest not to call this what it is – a neoliberal, pro-business, anti-labour ideology.

By contrast, the adoption of a developmental program would have given birth to a very different set of policies: address the unequal distribution of income and wealth; develop a wage-led growth path; make the satisfaction of basic needs the basis for eradicating the informal economy; end monopoly power in select sectors of the economy; give the state a central role in allocating resources including investment and, with respect to minerals, impose greater beneficiation of what is mined.

Apartheid was a system of general repression but the pain it inflicted on workers was particularly severe. Racial separation may appear to be the central plank of its ideology but it was in essence a system of super-exploitation, a pernicious way of extracting surplus value from farm labourers, miners and factory workers. This, and the colonial occupation that preceded it, is the historical root of the low-wage economy that we now live in. Data assembled here is sufficient evidence that the advent of political democracy has only perpetuated this low-wage system with its attendant poverty, unemployment and inequality. What is different now is the veneer of a “new social compact”, a system where labour and capital mediate their “antagonistic positions” with a neutral state.

Workers would do well to dismiss this approach to resolve the historical tensions that lie at the heart of capitalist production. For as long as this system prevails, they should continue to press for a greater share of the wealth they produce by using every available statutory instrument along with their labour power. Capital – labour relations have no prospect of improving given the logic of capitalism but even mainstream economists understand that one of the main reasons for South Africa’s laggard performance is weak aggregate demand. A wage-led growth path has the potential to stimulate demand and breathe life into ailing industries. But this alone will not be enough. It will need strong redistributive mechanisms to reduce the gaping inequality. And this will take nothing less than transforming the state.


Bhorat, Haroon, Karmen Naidoo, and Derek Yu. Trade Unions in an Emerging Economy. WIDER Working Paper. United Nations University, 2014. http://www.wider.unu.edu/publicatio...

Bhorat Haroon , Cheadle, Halton and Benjamin, Paul. The Cost of Doing Business and Labour regulation. International Labour Review 149(1) pp73 - 91. 2010.

Black, Anthony, and Hasson. “Capital Intensive Industrialisation and Comparative Advantage: Can South Africa Do Better in Labour Demanding Manufacturing?” University of Cape Town, South Africa, 2012. http://www.cssr.uct.ac.za/sites/css...

Buhlungu, Sakhela, and Malehoko Tshoaedi. COSATU’s Contested Legacy. HSRC Press, 2012.

Clarke, George R., Michale Ingram Habyarimana, David Kaplan, and Vijaya Ramachandran. An Assessment of the Investment Climate in South Africa. World Bank Publications, 2007.

COSATU. Findings of the COSATU Workers Survey 2012, 2012.

Di Paola, M., Pons-Vignon, N., 2013. Labour market restructuring in South Africa: low wages, high insecurity. Rev. Afr. Polit. Econ. 40, 628–638. doi:10.1080/03056244.2013.858432

Forslund, Dick. “Pampering of the Rich Too Costly.” Amandla!, February 25, 2013. http://www.amandla.org.za/component/content/article/1655-pampering-of-rich-too-costly--by-dick-forslund.

Galbraith, James K. “Kapital for the Twenty-First Century?,” 2014. http://www.dissentmagazine.org/article/kapital-for-the-twenty-first-century.
Hartford, Gavin. The Mining Industry Strike Wave: What Are The Causes And What Are The Solutions?, October 3, 2012. http://us-cdn.creamermedia.co.za/as...

Holmes, Thalia. “The Hidden Billions of SA’s Executive Elites.” Mail and Guardian Online, April 11, 2014.http://mg.co.za/article/2014-04-10-...

Kingdon, Geeta Gandhi, and John Knight. “How Flexible Are Wages in Response to Local Unemployment in South Africa?” Industrial and Labor Relations Review, 2006, 471–95.

Kingdon, Geeta, and John Knight. “What Have We Learnt about Unemployment from Microdatasets in South Africa?” Social Dynamics 27, no. 1 (2001): 79–95.

Morival, Elodie. “Top Incomes and Racial Inequality in South Africa Evidence from Tax Statistics and Household Surveys 1993 – 2008.” Master, Paris School of Economics, 2011. http://piketty.pse.ens.fr/fichiers/....

Onaran, Ozlem, and Giorgos Galanis. Is Aggregate Demand Wage-Led or Profit-Led? National. Conditions of Work and Employment Series. ILO, 2012.

Reddy, Niall. Wages, Profits and Unemployment in Post-Apartheid South Africa. AIDC Booklet. 2014. [Available Online]: http://aidc.org.za/media-room/publi...

Schroeder, Igsaan. Labour Brokers Bad for SA Workers, but Trade Unions Not Much Better. Video, 2012. http://sacsis.org.za/site/article/1230.

Steyn, Lisa. “SA Companies Are Raking in the Profits.” The M&G Online, June 14, 2013. http://mg.co.za/article/2013-06-14-...

Sunday Independent. “Fat Cats Get Fatter While Profits Fall,” April 27, 2014. http://www.iol.co.za/sundayindepend...

December 2014



[1Officially, according to StatsSA, “In order to be considered unemployed based on the official definition, three criteria must be met simultaneously: a person must be completely without work, currently available to work, and taking active steps to find work. The expanded definition excludes the requirement to have taken steps to find work.”

[2Kingdon and Knight, “What Have We Learnt about Unemployment from Microdatasets in South Africa?”

[3Kingdon and Knight, “How Flexible Are Wages in Response to Local Unemployment in South Africa?”

[4Ilan Strauss contributed to data work in this section. All errors are the authors.

[5The slight discontinuities between 1999 and 2000 can be partly attributed to improved methods of capturing informal workers with the shifts to the LFS.

[6The slight discontinuities between 1999 and 2000 can be partly attributed to improved methods of capturing informal workers with the shifts to the LFS.

[7Clarke et al., An assessment of the investment climate in South Africa.

[8Sunday Independent, “Fat Cats Get Fatter While Profits Fall”; Holmes, “The Hidden Billions of SA’s Executive Elites.”

[9Galbraith, “Kapital for the Twenty-First Century?”

[10Forslund, “Pampering of the Rich Too Costly.”

[11Morival, “Top Incomes and Racial Inequality in South Africa Evidence from Tax Statistics and Household Surveys 1993 – 2008.”

[12Schroeder, Labour Brokers Bad for SA Workers, but Trade Unions Not Much Better.

[13Cited in Di Paola and Pons-Vignon ’Labour market restructuring in South Africa: low wages, high insecurity’ (p. 633)


[15The National Economic, Development and Labour Council’s (NEDLAC)

[19COSATU, Findings of the COSATU Workers Survey 2012, 35.

[20Ibid, 24

[21Buhlungu and Tshoaedi, COSATU’s Contested Legacy, 21.

[22Ibid 22

[23Hartford, The Mining Industry Strike Wave: What Are The Causes And What Are The Solutions?.

[24Bhorat, Naidoo, and Yu, Trade Unions in an Emerging Economy.