Loane Sharp responds to Servaas van der Berg

Adcorp labour economist says StatsSA's own data points to hundreds of thousands of unemployed graduates

A response to Prof Servaas van der Berg

I am not going to get into a debate around the size of the informal sector in South Africa. This debate was had, in part on this website, with Prof Martin Wittenberg and Dr Andrew Kerr of the University of Cape Town, whose ignorance of the very best international research in this field was exposed in numerous forums.

The world's most reputed expert in estimating the size of the "shadow economy" (i.e. informal sector), Austrian Prof Friedrich Schneider, states: "Overregulation and labour costs in the official labour market are driving forces for the shadow economy. [Using the currency demand method,] an excessive increase in currency [...] is attributed to the rising tax burden and other factors leading people to work in the shadow economy. [...] The currency demand approach is one of the most commonly used methods. It has been applied to many OECD countries."

Prof Schneider and his colleagues estimated that South Africa's "shadow economy" was 9.0% of GDP in 1990. Updating that work for 2011, using the identical currency demand methodology that Prof Schneider himself uses, Adcorp finds that South Africa's informal sector is now 16% of GDP. An excellent summary of this body of research, together with criticisms of the currency demand method as well as comprehensive responses to those criticisms, is given in Schneider, F., and D.H. Enste (2000). "Shadow Economies: Size, Causes and Consequences", Journal of Economic Literature 38(1) 77-114.

I previously challenged Wittenberg and Kerr to publish their critiques of the currency demand method in an economic journal of similar stature as the journals which I cite, such as the Journal of Economic Literature, which is one of the world's foremost journals published by the American Economic Association. They have not done so, for the reason that their critiques are flimsy and would not stand up to peer review.

However, I am happy to have a debate around the reliability of Statistics SA data. This is an important debate. But Prof Servaas van der Berg of the University of Stellenbosch cannot be debated with. He claims that there are 55,000 unemployed graduates in South Africa. He is not looking at the same Quarterly Labour Force Survey (QLFS) as I am. In the QLFS of Q4 2012, out of 12,163 respondents who are counted as unemployed, 658 (or 5.4%) possess a tertiary or higher qualification, amounting to 364,932 of South Africa's 6.8 million unemployed ("broad" unemployment, i.e. not working or not seeking work) or 243,054 ("narrow" unemployment, i.e. not working and seeking work).

Even if you don't accept Adcorp's estimates and wholly believe Statistics SA's, you come up with 200,000-300,000 more unemployed graduates than van der Berg cites. This sort of weak conjecture with academic pretensions should be readily dismissed.

It is true that Adcorp inhabits two worlds: the official one, where we have consistently taken on the authorities for poor statistical work; and the unofficial one, where we have produced our own estimates and sense-checks of the official data. The disparity between these two worlds comes up, for example, when we are asked what South Africa's unemployment rate is.

On the one hand, we give the official answer, namely a narrow unemployment rate of 24.9% and a broad unemployment rate (i.e. including discouraged work-seekers) of 37.4%. On the other hand, we qualify the official answer by saying that there is an alternative view, promoted by Adcorp, that South Africa's unemployment rate is 9.7%. To claim, as van der Berg does, that these two worlds are inconsistent is correct: that is precisely the point.

Click here to sign up to receive our free daily headline email newsletter