Labour and unemployment in South Africa: towards a ‘grand bargain’

Reads 20,027
Ravi Kanbur, 7 October 2015
The problematics of the situation in South Africa are clear: high unemployment, high inequality and low growth, combined with a lack of consensus on what to do. It might be more fruitful to think in ‘grand bargain’ terms: a package of policies that are intended to balance opposing perspectives whose differences cannot be resolved through technical debate – and to set short-term political-economic imperatives against the longer time horizon needed for policy interventions to address deep structural legacies


The problematics of the situation in South Africa are clear: high unemployment, high inequality and low growth, combined with a lack of consensus on what to do – despite no shortage of actual and proposed policy interventions and no shortage of discussion of each.

Why is there such a lack of consensus? The matter may be purely a technical one to do with disagreements on the design of policies and interventions. Or it may relate to broader visions, or which political constraints are seen as binding and which not. Alternatively: could it relate to the fact that the key interventions will only be effective in 7-10-15 years’ time, while the political need is to deliver high growth, low unemployment and low inequality in 3-5-7 years?

My argument is that it might be more fruitful to think in ‘grand bargain’ terms: a package of policies that aim to balance opposing perspectives whose differences cannot be resolved through technical debate, and to set the needs of short-term political-economic imperatives against the longer time horizon needed for policy interventions that address deep structural legacies.

Structural initial conditions facing South Africa

Let us start with global trends that are faced by all countries. The first is the tendency of technology to displace unskilled labour by capital and by skilled labour. This shift in the structure of demand for the factors of production will, all other things being equal, increase the returns to capital relative to labour and the returns to skilled labour relative to unskilled labour. In turn, this increases income inequality through wage adjustments and/or unemployment.

The second trend is the great openness of the world economy to financial and direct investment flows. For a small country like South Africa, this is a global structural condition which has to be taken as a given. (A third global trend is that of climate change, with a time horizon of twenty to thirty years or more.)

Domestically the structural initial conditions in South Africa are well known. Many of these are the legacy of apartheid and the particular coalitions that struggled against apartheid:

  • large inequalities in financial, physical and human capital (education and health), especially along racial lines;
  • a particular spatial pattern of residence relative to place of work, leading to high reservation wages;
  • strong organised labour in the private and the public sectors – which can resist declines in real wages – combined with a small informal sector;
  • a significant export-oriented natural resource extraction sector with increasing capital intensity, being vulnerable to global commodity price trends; and
  • an economy relatively open to trade, investment and financial flows.

Policy dilemmas

The dilemmas in advancing the objectives of growth, employment and equity should be clear. Global technological trends, imported to South Africa because of its openness, are leading to the less intensive use of basic labour. The labour thus being displaced could be rehired by employers (a) if wages fell and/or (b) if the total demand for output rose, leading to more investment and thus more jobs. On the first, employers and policymakers face real wage resistance by organised labour. On the second, while being a small open economy helps to maintain demand, South African producers face competition from technically comparable producers in other countries who can rely on lower real wages to enhance their competitiveness. Moreover, macroeconomic strategies to address export competitiveness (e.g. ‘a mildly depreciated exchange rate’) could lead to a nominal-wage spiral as organised labour moves to protect real wages. These dilemmas are compounded by the relatively high reservation wage noted above.

Actually, some would point to labour market institutions as the key explanation of low investment and high unemployment, given that the private sector has the whole world to sell to and so does not face a demand constraint. But a low real-wage strategy does not seem to be available politically – at least for the 3-5-7-year time horizon – because of organised labour’s stance and a strong aversion to the perceived consequences for inequality.

South Africa seems to be in the following spiral. The juggernaut of technological change is displacing basic labour. Combined with a resistance to real wage reductions, there is high unemployment. With unemployment comes the pressure to expand employment in the public sector which, together with wage demands, leads to a rising public-sector wage bill. If the fiscal balance is protected through higher taxation, or if higher fiscal deficits elicit tight monetary policy, employment in the private sector is hit again – adding to the way in which global forces contribute to rising unemployment. The spiral continues.

A plethora of specific polices and interventions have been proposed and implemented, including:

  • greater expenditure on public education, health, housing, services and transfers;
  • public employment schemes;
  • skills development and skills matching services;
  • wage subsidies to employers;
  • minimum wages;
  • support for very small-scale enterprises in the informal sector;
  • export and investment promotion for enterprises in the formal sector;
  • public investment to counter infrastructure failures that hold back private investment; and
  • fiscal deficit targets and monetary policy that targets inflation.

These diverse policies and interventions each have their own rationale and there are vigorous debates on their design, implementation and efficacy.

These interventions are not necessarily consistent. A range of capital subsidies favour the use of capital over labour. Wage subsidies to employers counter high wages in the organised sector at the same time that a minimum wage would put a floor under wages. Public employment schemes can, at best, be a temporary device, lest those so employed become a permanent burden on the public purse. And public expenditure depends on public revenue, which will grow sustainably only when the economy grows through greater investment.

These policy inconsistencies may simply be due to oversight or a lack of policy coordination. Or they may actually reflect conflicting visions and attempts to reconcile many constraints. As a result, policy often undoes with one hand what it has just done with the other.

With specific reference to labour and employment, there are two strong competing perspectives. One is the standard labour-economics view that, if only the real wage would adjust sufficiently downwards, unemployed labour would be absorbed into employment. A value of 0.7 for the conventional wage elasticity of employment is marshalled in evidence – i.e. a 10% fall in the minimum wage would lead to a 7% rise in employment. The counter to this perspective is equally well rehearsed. Halving the unemployment rate would require around a 40% decline in the real wage, which is beyond the realms of political feasibility, let alone its implications for inequality between labour and capital.

(One should note that the research underlying the 0.7 number is now quite dated and the current elasticity may be smaller, given the pace of technical change. Some support for the effect of technical change is provided by the low output elasticity of employment: it is around 0.5, so that a 2% increase in output would be needed for every 1% increase in employment. Furthermore, in the short run the elasticity of both wages and output are likely to be much smaller.)

With this background, we can pose the following types of questions:

  1. How can each policy or intervention be improved in its own terms, given the specific objectives of each?
  2. Can we identify any inconsistencies between pairs of policies so that they do not counteract each other?
  3. Which of the policies provides quantitatively the biggest bang for growth, employment and equity?
  4. What might be the contours of a ‘grand bargain’ between competing perspectives that would advance growth, employment and equity?
  5. What specific types of policy could fill out such contours of a grand bargain?

There has been a fair amount of discussion of question 1, but progressively less of each subsequent question. Here I focus on questions 4 and 5 as perhaps the most useful at the moment.

A useful comparator: Brazil

Are there country experiences that might be relevant to the South African dilemma in terms of broad strategies that have delivered the desired outcomes in a situation similar to that facing South Africa?  Of course there is no direct comparator, given South Africa’s unique history. But even finding a middle-income country with at least some features structurally similar to those of South Africa is difficult:

  • No other sub-Saharan African country approximates the combination of characteristics in terms of size, income, inequality, development, openness, urbanization, market economy, formality, labour organisation and political liberation.
  • Middle East and North African economies are only now entering a democratic phase.
  • The transition economies of Eastern and Central Europe have had their transition to democracy and markets, but they started off with much higher levels of equality and human development – many are at EU levels of development (and are EU members). None has faced the racial divides faced by South Africa.

Brazil is a country from which South Africa has much to learn. It is not South Africa by a long shot, including having a much larger population. But it has many common structural features. It is a middle-income country and has high levels of inequality. Economic differences based on racial origins are well documented and are part of the discourse. Brazilians threw off the dictator’s yoke and have moved to social democratic government. Brazil is urbanized and formalized to a significant extent – and labour organisations are strong (the government for the last ten years has been formed by the Workers’ Party).

Despite Brazil’s structural similarity to South Africa, the main comparison lies in a difference. In the last fifteen years, Brazil has had high growth rates combined with declining inequality. (This followed various phases in which growth was high but with rising inequality (1960s and 1970s) and when growth was low with stable inequality, i.e. the 1980s.)

Brazil’s Gini coefficient was 0.63 at its peak, but has been declining since the late 1990s. Between 1998 and 2009 the Gini declined from 0.59 to 0.54. This decline has a huge quantitative significance: ‘two thirds of the decline in extreme poverty can be attributed to the reduction in inequality. For the same reduction in extreme poverty, Brazil’s overall per capita income would have needed to grow an extra 4 percentage points per year’ (Lustig, Lopez-Calva & Ortiz 2011).[1]

What accounts for the unprecedented reduction in inequality? According to Lustig, Lopez-Calva & Ortiz (2011), it is mainly the result of three factors:

  • decreasing wage differentials by educational level and reductions in the inequality in education;
  • increasing spatial and sectoral integration of labour markets, in particular among metropolitan and non-metropolitan areas; and
  • larger and better targeted non-contributory government transfers.

(Raising the minimum wage played a role via 1 and 3.)

They further report on the significant role of education. The 1990s was marked by an accelerated expansion of basic education in Brazil. As a result, the degree of inequality in education declined markedly. At the same time, the skill premium – the differences in the returns to education, particularly for those with secondary and higher education vis-à-vis workers without schooling or with incomplete primary schooling – also declined. The combined effect of these two factors – a fall in the inequality of education and a fall in the steepness of the returns to education – explains half of the decline in the inequality of labour earnings and almost 30% of the decline in household per capita income inequality.

It is remarkable that the wage premium declined in the face of the global trends in technical change alluded to earlier. This was the result of the strong supply-side effect of the accelerated expansion of basic education in Brazil (Lustig, Lopez-Calva & Ortiz  2011).

The strong backdrop to these changes in the labour arena was a policy of macroeconomic stability with firm control of fiscal deficits. This was extremely important, given Brazil’s history of high inflation, with spirals of price increases being chased by nominal wage increases to maintain real wages – and with nominal exchange rate depreciation adding to the explosive mixture.

The prominence of the macroeconomic stability objective was a distinctive feature of both the social democratic party (Cardoso’s presidency) and the workers’ party’s (Lula’s presidency). With this key component of an implicit grand bargain agreed upon and helped by the long period of commodity price increases up to 2008, the other components were put in place: the expansion of basic education, minimum wages and conditional cash transfers.

It is very important to note the time lags: it was a decade before the effects of the expansion of basic education on inequality were felt. The South African political imperative is to address the unemployment problem over the 3-5-7 year time horizon – and to do so without expanding employment in the public sector to fiscally unsustainable levels.

Thoughts for South Africa

The Brazilian experience can at best be only a stimulus to South African thinking on its core policy dilemma. Here are some thoughts on possible contours of a grand settlement:

  • Fiscal rectitude will have to be a key component of the package. We can discuss the detailed quantitative magnitudes involved, and it could be argued that the devil is indeed in the details, but this element is essential.
  • Expanding public investment in education and health (i.e. human capital), as the basis of raising productivity and reducing inequality, is essential. Again, the detail matters. But the benefits of this will only come slowly, over a decade or more. (Of course, the same is true of other medium-term tracks such as industrial policy.)
  • In the short run, competing in the world economy on a low real-wage track is not a feasible and in some quarters not a desirable policy option.
  • Publicly provided employment can at best be a temporary insurance intervention. The bulk of employment will have to come from the private sector.
  • Given these constraints, the focus in the short run – the feasible and desirable trackmay have to be to reduce non-labour cost of private sector employment. It could provide quick gains on the employment front.

The dilemma in South African policy can perhaps be crystallized as follows. How can unemployment be reduced significantly in the short run (a time horizon of 3-5-7 years) without lowering the real wage and without increasing the public-sector wage bill?

A focus on the non-labour cost of employment for the private sector opens up a policy agenda which might be less contentious, and more amenable to consensus than one focused on reducing labour cost. It also highlights how little we know in detail about the impact of non-labour cost on employment. A conventional response might be that the non-labour cost of employment account for less than 30% of total cost, therefore the impact is unlikely to be large. But this could potentially confuse average cost with marginal cost.

In addition, we can think of non-labour cost in a broad rather than a narrow sense and so bring in a wider range of possible policy responses. Here are some examples, each of which needs further study to quantify them:[2]

  • public infrastructure improvements for enterprises in their current location;
  • infrastructure improvements and specific financial inducements for enterprises to move to high unemployment locations;
  • steps to reduce the high reservation wage induced by the high cost of transportation to work;
  • steps to reduce monopoly and concentration on the product side in order to expand output and thus employment; and
  • the reduction of the regulatory and other constraints on the expansion of activity in small-scale informal activities.


I think the elements of a ‘grand bargain’ may have to comprise a move away from a ‘reduce-labour-cost’ approach to employment policy in return for an agreement on fiscal rectitude and the non-expansion of the public-sector wage bill. Within the fiscal constraints, the medium-term strategy would be to build up human capital through investment in education and health, while the short-term strategy, over the crucial 3-5-7 year time horizon, would be to reduce the non-labour cost of private-sector employment.

There is plenty of room for debate on the specifics, of course, but this combination may move us away from a divisive discourse towards an assessment of the details of each instrument in the context of a potential overall consensus.

* This is an abridged and edited version of a DPRU Policy Brief published by the Development Policy Research Unit, University of Cape Town, August 2015.


Barros R, de Carvalho M, Franco S. & Mendonça R. (2009). Markets, the State and the Dynamics of Inequality: Brazil’s Case Study, Research for Public Policy, Inclusive Development, ID-14-2009, RBLAC-UNDP, New York.

----------- (2010). Markets, the state and the dynamics of inequality in Brazil. In Luis F. López Calva & Nora Lustig (eds.), Declining Inequality in Latin America: A Decade of Progress? Chapter 6. Washing­ton DC: Brookings Institution and UNDP.

Lustig N, Lopez-Calva LF & Ortiz-Juarez E. (2011). The Decline in Inequality in Latin America: How Much, Since When and Why. Working Paper 1118, Department of Economics, Tulane University.

[1] Based on Barros, de Carvalho, Franco & Mendonça (2009 and 2010).

[2] These and other policy proposals are also to be found in Bhorat, Hirsch, Kanbur and Ncube (Editors) 2014. The Oxford Companion to the Economics of South Africa, Oxford University Press

Download article

PDF icon Kanbur 2015 SA unemployment and grand bargain - FINAL.pdf
Post a commentary

This comment facility is intended for considered commentaries to stimulate substantive debate. Comments may be screened by an editor before they appear online. To comment one must be registered and logged in.

Symptoms or root causes?

There is no doubt that this article identifies many of the macro factors impacting on South Africa’s potential to develop its economy and bridge the inequality gap. However, I think there is a meta level factor that is either being ignored because of political sensitivity, fear of alienating Official Development Assistance (ODA), or, simply not being on the radar.

This factor is an apparent discord in political and economic ideology within the ANC. Influential elements in the ANC are undoubtedly capitalist but the rhetoric is consistently socialist, if not Marxist. This discord seems attributable to the influence of the South African Communist Party (SACP) whose mission is state control of all South Africa’s factors of production, i.e. land, raw material, and labour.

This power that is being wielded by the SACP resides in their success in entrenched themselves within all the alliance partners of the ANC and in establishing themselves as the intellectual and organisational backbone of the ANC as far back as the 1960’s. (See External Mission: The ANC in Exile 1960 to 1990 – Stephen Ellis; Oxford University Press). The manifestation of this mission can be found in ever increasing centralisation of planning and control; dismissal of any pretence of empowerment and subsidiarity; a myopic view of development achievements; increasingly “bigger government”, to an extent that would make Keynesian disciples eyes water; and, an apparent determination to reign in all social advances that are being made by the private sector, such as education, security, health care, etc.

I believe that a strong case can be made that the ANC’s mission to eradicate apartheid has been achieved. Their role as a revolutionary organisation is over, and they are in power, they are the government, and their only purpose now is to create an environment where all the resources of the country are effectively mobilised to create the type of growth that is needed to create jobs and reduce inequality. This means learning from what has, is, working and replicate and empower rather than impede and demonise. If, on the other hand, the ANC, SACP, and COSATU are unified in a socialist or Marxist vision for South Africa then the sooner that is made clear the sooner all concerned can all make informed decisions and move forward. 

This comment facility is intended for considered commentaries to stimulate substantive debate. Comments may be screened by an editor before they appear online. Please view "Submitting a commentary" for more information.