Why Mike Schussler is wrong about inequality

Schussler’s article has three central thrusts.

First, he argues that South Africa is not as unequal as popularly thought.

Second, Schussler asserts that ‘the major cause of inequality is not the pay gap, but rather the low number of people earning wage income, that is unemployment is to blame.

Third, he claims that the cause of high unemployment is “militant labour organisations” making “demands for ever increasing wages”. Remarkably, the conclusion is that increasing wages for the working poor increases inequality.

The claim that South Africa is not as unequal as often believed relies on manipulations of the Gini Coefficient, a standardised global measure of inequality ranging from 0 to 1, with 0 the most equal and 1 being the most unequal. Schussler argues that if you take account of social grants, education, health provision, and “the highly progressive nature of our tax system” the world-leading 0.69 Gini Coefficient drops to 0.47 – in line with the UK and USA. When isolating those in the “small formal sector” the Gini coefficient is as low as 0.28, close to Germany or Norway, some of the most equal countries in the world.

Frankly, this proposition is ridiculous. Someone best tell the almost 7 million people living in shacks in over 2,700 informal settlements that despite the differences between their lifestyle and those with swimming pools and tennis courts in Bishopscourt and Sandton we’re actually a very equal society.

The findings of the leading studies on inequality in South Africa are unequivocal: inequality is extremely high and increasing and wage inequality is a key culprit.

The Data

Some sophisticated data sets shed light on inequality, the most recent of which is the Presidency and United Nations sponsored National Income Dynamics Study (NIDS), the first national household panel study in South Africa, conducted by the Southern Africa Labour and Development Research Unit (SALDRU) based at the University of Cape Town’s School of Economics (with comparisons made to datasets from 1993, 2000 and 2005).

More importantly, the NIDS data and the conclusions below are used by the world’s leading research organisations and statistic depositories when analysing inequality in South Africa. These include the World Bank, Organisation for Economic Cooperation and Development (OECD) (here and here), International Labour Organisation (ILO) and Luxembourg Income Study (LIS), as well as by the South African National Planning Commission (including in the National Development Plan) and the Statistician-General of South Africa.

There are flaws in all data sets (see discussion on NIDS data and review of all SA datasets) but NIDS is by far the best we have. Interestingly, the researchers believe that the survey underestimates inequality because the richest households refuse to answer income-related questions. In other countries this is remedied by using tax data, but this is not publicly available in South Africa.

Unfortunately Schussler appears not to use the NIDS studies although this is difficult to confirm as he provides only vague references for almost all of his data and few hyperlinks (“UBS bank reports”, “a KZN university report” and so on). He also appears to make cross-country comparisons with incomparable data (discussed below).

How High is Inequality?

Leibbrandt et al. explain that there is consensus within the literature that “both aggregate inequality and inequality within each race group has continued to increase through the 1990s and into the 2000s”. The overall after-tax Gini coefficient rose from 0.66 in 1993 to 0.7, making it one of the highest in the world.

Schussler (presumably on the basis of a 2009 Van der Berg paper) adjusts the Gini coefficient downwards to take account of progressive tax redistribution, social grants and social services. The methodology used is unclear and no standard international method exists for adjusting Ginis on the basis of the provision of social services. Before proceeding Van der Berg acknowledges this to be an incredibly fraught process and the data to be insufficient. (Van der Berg does not use NIDS data as the data was not yet available when the paper was drafted).

The NIDS studies’ Gini of 0.7 already accounts for social grants and taxation (the before-tax “market Gini” is actually 0.75), meaning that part of Schussler’s downwards adjustments amount to “double counting” these factors.

The authors do acknowledge that income inequality does not give the full picture and that the poor have benefited from state-funded asset provision, such as houses and access to services.

There is no doubt that social services have been pro-poor and would reduce inequality if considered. However, it is incorrect to compare Gini coefficients adjusted downwards for tax and social transfers in South Africa with other countries’ unadjusted coefficients. Schussler is not comparing apples with apples. Putting aside the problems with carrying out such an exercise, you would have to adjust the gini coefficients in the comparator countries by taking their social welfare measures into account too, using OECD estimates. Schussler’s 0.47 is well above the OECD 2010 average of 0.3, with the US and UK —to which Schussler claims SA is “close to”— at 0.34 and 0.38.

What is Responsible for Inequality?

The studies separate the impact of the labour market (income derived from wages), state transfers (income from grants), remittances (income sent home by a household member working elsewhere) and capital income (income from dividends, interest, rent, pensions and the like).

State grants have helped to decrease poverty but have had “virtually no effect on overall inequality”. This is because grants have managed to shift some people from the lowest income bracket into the lower-middle sections but not close the gap between the rich and poor. Remittances have decreased significantly and actually slightly increased inequality, possibly because they have become highly correlated with labour market developments. Capital income is the most unequal and contributes to inequality but this effect is somewhat limited because it makes up only 8% of total income.

This means that “the labour market is the driving force behind aggregate inequality in the country”; between 1993 and 2008 wage income (including self-employment income) accounted for 85% to 90% of inequality. This is because around 70% of households receive some wage income, and overall income is closely tied to wage income.

What is to “blame” for this? This is tackled head-on by separating the contribution to inequality caused by “wage gaps” –the difference in income between those earning– and the contribution to inequality made by some not earning at all, i.e. unemployment.

Using a well-recognised model, inequality is decomposed between these two factors. The data shows that 62% of inequality is accounted for by differences in wages, 38% by unemployment. Unemployment plays an important role in inequality but not the leading one.

This fits with what we know about wage inequality, as income in South Africa has increasingly been concentrated at the top. As Leibbardt et al. note in the respected journal Development Southern Africa:

In fact, in 2008 the wealthiest 10% accounted for 58% of total income. The richest 5% had a 43% share of total income, up from about 38% in 1993. The cumulative share of income accruing to the poorest 50% dropped from 8.32% in 1993 to 7.79% in 2008.

Not only have the poorest decreased their share of income but real wages for these groups have actually declined (see below).

The evidence contradicts Schussler’s arguments. The 0.7 Gini coefficient already takes account of tax and social grants. Social grants have a negligible impact and wage inequality plays a greater role than unemployment. It turns out that South Africa is just as unequal as we all thought and growing wage inequality matters greatly.

Are High Wages Responsible for Inequality?

Are greedy unions that make ever increasing wage demands retarding employment? Once again the evidence suggests otherwise.

First, unions have not succeeded at achieving “ever increasing wages”. Between 1997 and 2008 only the wealthiest 20% in South Africa enjoyed any growth in real wages. All other groups suffered a decline. Real wages for the poorest 10% halved, and rose by 23.5% for the richest 10%.

Second, there is not consensus among economists that lower wages will lead to substantially greater employment. Increasing employment, something we desperately need to do, depends on levels of investment, and real investment depends on a range of factors including infrastructure, available technology, logistics, geographical location, natural resources, price inputs, a skilled workforce, reasonable borrowing costs and so on, not just wage levels.

Third, the same studies show clearly that “having a job on its own is not a guarantee that a household will move into the top deciles”, that is, expanding employment does not automatically reduce inequality. The quality of employment, including the wage level, matters.

Schussler’s assertion that wage increases retard employment and therefore increase inequality should be viewed, like the rest of the unsubstantiated claims he makes, with scepticism.