It is always salutary to be reminded just how dire are the economic circumstances of the average South African and how slowly their economic conditions have been improving. Over 51%, some 29,733,210 of our people, live on less than R1,036.07 a month. These and many other shocking statistics are reported by the panel of experts appointed to recommend the level of a national minimum wage and on a process for its effective implementation, in their report to Deputy President Cyril Ramaphosa.
The panel has no doubts about the helpfulness of a national minimum wage in principle — only reservations about practice. To quote selectively from its substantial 128-pages report:
"On its own it will not solve all of the challenges we face, but it is an implementable policy which is designed to have a measurable and concrete benefit on the poor. The minimum wage is therefore seen as one of the tools to close the wage gap, including between the genders, and thereby to overcome poverty.
"Furthermore, under the correct conditions and at the correct wage level, it is possible for minimum wage policies to contribute to improving economic growth.… Given that the national minimum wage is essentially a policy to help the poor, it is generally accepted that exemptions and exclusions should be kept to an absolute minimum."
Striking a balance
The panel was required by the social partners in the National Economic Development and Labour Council (Nedlac), who agreed to a national minimum wage, to recommend an appropriate level for it. Since it recognised a relationship between wages and employment, the national minimum wage had to strike a balance between the effects of increasing wages to a higher prescribed minimum level and its consequences for unemployment — of which SA already has a great abundance. Some 26% of the labour force, those in work or looking for work, are unemployed, while many more potential workers have been discouraged from looking for work and have fallen out of the labour force. Adding them to the work force would imply a more broadly defined national unemployment rate well into the 30%-plus range.
The panel recommended a national minimum wage of R3,500 a month, or R20 an hour, to be phased in by 2020, with 90% of the national minimum wage to be applied in agriculture and 70% in domestic service provided to private homes.
It also recommended annual reviews of the national minimum wage and a gradual move to uniformity across all sectors of the economy.
It used a so-called computable general equilibrium model of the South African labour market (as described and developed by Prof Harron Bhorat of UCT) that included assumptions (not predictions based on past performance) about the trade-offs between percentage increases in minimum wages and percentage reductions in employment (the relevant employment elasticities, in economic speak).
The aggregate employment losses were estimated as between 100,000 and 900,000 jobs lost in exchange for the recommended national minimum wage.
Clearly in the view of the panel, the national minimum wage had to be set high enough to make its contribution to poverty relief and yet low enough to be able to treat the extra unemployment as the acceptable price to the panel members of them doing such good for society — another argument for breaking eggs to make omelettes from those unlikely to be harmed by the action and who might even benefit from helping to give effect to a new dispensation.
However, the panel did qualify its judgment. It noted correctly, "that there is no research or data that can accurately predict the outcome of any policy intervention. It is for this reason that strong emphasis has been placed on the need for good solid research to support the work of the national minimum wage institution into the future. Any future changes to the level of R3,500/R20 an hour should be based on solid evidence of the impact of the national minimum wage."
Would it be unkind to recognise that this would also be more grist for the economist’s mill?
Minimum wages and employment
It would have been helpful had the panel used the report to explain more fully why employment offers are negatively related to the wages, or rather employment benefits provided by employers in exchange for hours worked. The relationship is much less self-evident than the panel may have presumed it to be — especially to members and leaders of trade unions, who are inclined to attribute wage differences much more to political forces and bargaining power and even to race, than to the differences in skills and therefore of the contributions to output made by the well and poorly paid.
And they are very inclined to believe that the wage gaps can be easily closed, with little consequence for economic growth, by taking more from the well paid and giving to the poor.
The panel is well aware of an obviously important and highly consistent relationship to be observed in the SA labour market between measures of skills and wages earned. The relationship between incomes and employment is just as statistically significant. The lowest-paid South Africans have the highest rates of unemployment. The full income and employment details are shown in the table below.
Other reports from Stats SA have demonstrated the links between educational attainments and income and employment.
It may be seen in the table that of the 16,306,000 people in quintile 1, 31.2% of the population with the lowest share of income, only 15.9% are employed, 25% strictly unemployed and the broad unemployment rate of this group is estimated as 65.8%. The average monthly wage of those employed in quintile 1 is only R1,017.
The second-poorest quintile accounts for a further 24.6% of the population, has a lower unemployment rate, a much higher participation rate in the economy and an average monthly wage of R1,707 — a large improvement but still a very low average wage.
The top income quintiles present very differently. Unemployment rates are much lower, and participation rates and average incomes are much higher and well above the recommended national minimum wage. Though it should be noted that the average monthly wage of those employed who fall into quintile 3, at R2,651, is still well below the R3,500 recommended national minimum wage.
The implications of a minimum wage far above average wages
Is there any precedent for a national minimum wage or a sectoral minimum wage determination to be set so far above average earnings? If so, what have been the consequences for employment?
Put another way, is there reason, given the facts of the labour market, to think the employment elasticities in SA are a lot more negative than the range of assumptions considered by the models?
Time will tell much more about the consequences of the recommended national minimum wage, as the panel complacently assumes, and adjustments can then be made to the model and the recommendations. But who will care for the unemployed and their dependents in the meanwhile?
It is very difficult to understand why the panel should believe the national minimum wage can be helpful to the poor of SA or reduce inequality. Because fundamentally the poorest South Africans — those in first and second quintiles — are mostly not employed. And when employed, they command wages far below those of the recommended national minimum wage that still leave them objectively poor.
The recommended national minimum wage will surely make it even more difficult for them to find work that might, for some, especially young workers, prove a path out of poverty. Thus the national minimum wage is very likely to increase further the unemployment of low-skilled potential workers in SA and to widen the gap between the average incomes of the high earners and the low earners, mostly no earners, of the population.
The poor of SA deserve better opportunities to work, and the opportunity for their children to acquire the education and skills that would help them qualify for and find well-paid work. They do not need further interference in their search for work.
Employers’ response will confuse the evidence
To complicate the numerical outcomes to be observed in due course, employers will make structural adjustments in response to higher minimum wages, such as relying more on mechanisation and automation — requiring more carefully selected and skilled employees. The forces that substitute capital for labour, especially less skilled labour, are already well at work in the economy.
Employers will also offer fewer hours of work and cut other employment benefits, such as food and accommodation and contributions to pension and medical aid. The panel, in its money-only determination, refuses to recognise the cost to employers and value to employees of these benefits.
Employers’ reactions have been seen in wage adjudications in agriculture: fewer workers were employed permanently, less accommodation was offered on farms, and workers incurred transport costs.
There is also bound to be less compliance with the law, given the availability of cheaper labour and additional employment offered to illegal immigrants.
Opportunity costs
A further observation of the inconvenient and uncomfortable truths of the South African labour market is that the supply and demand for labour are very well matched for the well paid and very poorly matched for the low paid.
A very high unemployment rate is surely evidence of wage levels that are too high, rather than too low, and undermines the important purpose of providing work for those who would wish to work at prevailing wages.
These are not considerations that receive much attention from the panel. Why do these structural forces that discourage employment not apply to the most expensive of workers in SA, who are so readily employed?
It is to be conceded that employment at low wages for those with limited skills cannot overcome the poverty of the working poor. But then what can? Other than them acquiring the valuable skills that are in short supply and well worth hiring.
Wishful thinking — or waving magic wands in the form of unaffordable minimum wages — will not solve their problem.
Unemployment makes their condition more onerous and denies them the employment and low wage benefits they would be willing to accept — a willingness demonstrated by their seeking work. And being unemployed prevents the potential worker from acquiring skills on the job and the opportunity to demonstrate their capabilities that add to their employment credentials.
These opportunities are particularly important to young, unskilled, entry-level workers whose unemployment rates are regrettably but understandably well above average unemployment rates, as is well recognised by the panel.
The panel might have sought an explanation of the high rates of unemployment of low-income South Africans in the structural impediments to their employment. It is not as if minimum wages have not been tried in SA. They are widely practised and have surely had their effect on the employment of the lowest paid and least skilled.
Regulatory barriers to employment
There are in fact 124 separate such sectoral minimum wage determinations. They cover approximately 5-million workers and 33% of those employed — leaving only 35% of workers uncovered including presumably many of the better paid who also lack union representation.
The lowest such monthly determinations in 2015 ranged from R1,813 for domestic workers to R2,844 for contract cleaners in the lowest grades.
The highest sectoral minimum determinations, for more skilled work, were R6,155 a month for workers in private security and R6,506 a month in retail and wholesale businesses.
The newly fashioned national minimum wage is intended to remove all this administrative complexity — and presumably also the possibility of recognising very different labour market supply and demand conditions that may apply in the different sectors and regions of the economy.
Participants in specific labour markets, unencumbered by regulations, would be much better informed about these conditions than even diligent officials, to the advantage of workers and their employers.
The case for leaving the determination of an employment contract to willing buyers and sellers of labour does not get any hearing from the panel. The collective bargaining process, which can easily be shown to protect the established interests of employees and their employers — the insiders — at the expense of the employment opportunities of the outsiders, receives nothing but uncritical approval from the panel, and an appeal for the wider application of collective bargaining arrangements.
The influence of welfare on employment
Nor did the influence of SA’s extensive welfare system on poverty and employment receive much more than perfunctory and rather condescending attention from the panel, which says in its report:
"While wages are low relative to living levels, there are arguably some offsetting effects from the social wage spending by government. About 35% of SA’s budget is spent on programmes targeted at the poor, including free basic education, healthcare, water and electricity, and income support grants for children and the elderly."
One can only wonder how panel members could have argued that social wage spending could not have some offsetting effects on consumption and, more arguably perhaps, on employment via the willingness to supply labour services at low hourly rates on offer.
They may, as did the Davis Committee on Tax Reforms, have referred to a report the World Bank issued in November 2014, in which it said:
"But while incomes earned in SA may well be the most unequally distributed in the world, the distribution of expenditure is much less unequal. The World Bank shows, in a recent study, that SA does more to redistribute income in cash and kind to the poor than its developing economy peers with similar average incomes, Armenia, Brazil, Bolivia, Costa Rica, El Salvador, Ethiopia, Guatemala, Indonesia, Mexico, Peru and Uruguay."
As this study also reports:
"SA ranks as one of the most unequal countries of CEQ (commitment to equality methodologies, applied by official statisticians in income measurement) participant countries, if not among all middle-income countries, given its Gini coefficient of 0.69. The proportion of the population living in poverty, at 33.4% measured by the international benchmark of $2.50 a day (purchasing power parity, PPP, adjusted) — is also higher than in many other middle-income countries with similar levels of GNI per capita. For example, the poverty rate is 11% in Brazil and 4% in Costa Rica."
To quote further from the World Bank report:
"Briefly, this Update has two main findings. First, the burden of taxes falls on the richest in SA, and social spending results in sizable increases in the incomes of the poor. In other words, the tax and social spending system is overall progressive.
"Second, fiscal policy in SA achieves appreciable reductions in poverty and income inequality, and these reductions are in fact the largest achieved in the emerging market countries that have so far been included in the CEQ.
"Yet despite fiscal policy being both progressive and equalising, the levels of poverty and inequality that remain are unacceptably high. SA is currently grappling with slowing economic growth, a high fiscal deficit, and a rising debt burden. In this context, addressing the twin challenges of poverty and inequality will require not only much-improved quality and efficiency of public services but also higher and more-inclusive economic growth to help create jobs and lift incomes."
These income transfers and benefits in kind may influence the willingness to supply labour services at prevailing wages — especially when wages on offer are very low. By providing an alternative source of benefits, welfare raises the reservation wage — the wage at which it makes good sense to work or to seek work, which may well be physically demanding and less than enjoyable for its own sake.
The panel might have paid much more attention to the supply side of SA’s labour market to help explain low rates of labour force participation — as well as to help explain why immigrants from Africa are much more likely to be employed at market-related wages. They clearly have a lower reservation wage, given their limited access to South African welfare and the dependence their families left behind may have on cash remitted to them.
Growth and jobs: ignoring the evidence
The panel remarks somewhat self-evidently that:
"An additional problem faced by the country is that there is evidence that the growth in the demand for labour in SA has not been sufficient to keep up with the much larger growth in labour supply."
The panel quotes, with seeming approval, a study that apparently shows growth and job creation are not well correlated. To quote the panel:
"Recent empirical work by Mkhize (2016) finds that the economy’s capital intensity undermines its ability to generate jobs in times of economic growth. He finds that, in the long run, growth and job creation are not correlated, although there is some sectoral variation. This points to the broader economic policy challenge facing SA, which is that there are structural barriers that exacerbate unemployment, the solutions to which require more than economic growth."
A surprising conclusion, it might be thought, given the fact that in the developed world, incomes, GDP, population and the size of the labour force have grown together — as indeed they have in SA, as our own work has shown (see below) — though work on the relationship in SA between GDP and employment is complicated by the absence of an official continuous long time series of numbers employed.
Reserve Bank employment series data used to calculate the figure above, dates only from 1967 and refers to employment outside of agriculture and to formal employment only. The numbers employed in these categories amounted to 9.17-million employees in 2015.
The predictions of this model are compared with the employment outcomes below. It shows the model overestimates employment after 1990 and underestimates employment more recently.
Using earlier sources of information on employment before 1967, we have estimated an employment series going back to 1940. The predicted relationship with GDP (available from 1946) is shown below.
The relationship between GDP and employment over this longer period remains highly significant, with the GDP influence a somewhat weaker one.
The evidence strongly suggests the relationship between GDP and employment in SA remains as highly significant as economic theory and history would predict.
Though clearly there is more to employment than GDP, any increase in GDP growth can be confidently predicted to increase employment — as it did through the boom years of 2004-08, when GDP growth averaged above 5% a year.
The Mkhize study to which the panel refers is in reality nothing like a long-run study. The data to which it refers extends only to 2000. Furthermore, the study reveals a statistically significant positive relationship between output (GDP) and employment. The output influence on employment is particularly strong for the service sectors.
Clearly the relationship between population growth and growth in employment has weakened over the long run, as evidenced by a declining rate of participation of the adult population in the labour force.
The reasons for this may include a technological bias in favour of capital substituted for labour. But further reasons worthy of serious consideration, in addition to a welfare-enhanced higher reservation wage, are surely the plethora of interventions in the labour market, including the role played by trade unions willing and able, given regulatory advantages, to trade off lower employment for improved benefits for those who retain their jobs.
The recommended national minimum wage represents more of the same lack of faith in market forces that encourages further regulation of the labour market in SA, rather than a very different recommendation to introduce less interference in the market place to generate faster growth and employment.
It represents another example of economists, as are the governments they usually serve, being part of the economic problem rather than the solution.
• Kantor is chief economist and strategist at Investec Wealth & Investment. The views expressed in this article are those of the author and may not necessarily represent those of Investec Wealth & Investment.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.