Thumb-sucked wage statistics not useful in jobs row
As the debate in South Africa rages over the poorest workers allegedly being overpaid, the Occupy Wall Street movement, now replicated globally, expresses growing outrage at the “1%” of people whose wealth allows them to dominate the world.
What would probably further outrage the protesters is that the bankers and other corporate heads among these superrich “one percenters” sometimes manage to masquerade as ordinary workers. The magic used to achieve this gives a wholly new meaning to “lies, damned lies and statistics”.
Official gross domestic product statistics in many countries, including South Africa, recognise only two kinds of income: wages and profits. Everyone working on an employment contract (including the American hedge-fund manager who made $4-billion – yes, billion – in a single year) are listed as employees and their pay is registered as wages in the national accounts. The consequences of this statistical quirk are enormous.
One might expect the economists, who are household names because the media defers to them as “experts”, to alert us to this peculiarity.
Worse than not telling us, however, is that they use sleight of hand to promote their own far from neutral agenda. We are inundated by “research” supposedly showing that South African workers are not only grossly overpaid (despite their officially recognised poverty wages), but that this excessive pay is responsible for mass unemployment.
What they do not say is that the “average” pay they quote – R13 200 a month – includes the pay of all managers and executives. Also included are the earnings of all employers who pay themselves a salary, as well as the self-employed (including lawyers, cheap at R40 000 a day). These distortions reflecting South Africa’s notorious inequality are of such a magnitude that they make meaningless any “average” wage. If five out of 100 people earn R1000 and 95 earn R5, the average is R54.75. This means that 95 out of the 100 people have had their wages artificially increased by almost 11 times. It is magic.
Yet this statistical socialism, based on the premise of an equal sharing of wealth by everyone, has not stopped the business press from attacking real workers with lurid front-page headlines such as “Unskilled workers ‘paid too much’ in SA” (Business Day).
Ideology probably explains why many economists fail to correct these distortions. Much of the information required for such a correction is available. This has not stopped the media from attacking real workers, as opposed to the fictional ones of “averages”, for being the cause of so much of South Africa’s economic problems.
Section 27(1) of our Employment Equity Act 1997, for example, states that every employer (other than small ones) “must submit a statement … on the remuneration and benefits received in each occupational category and level of that of that employer’s workforce”. The purpose of this 15-year-old section was to reduce income differentials because, even then, they were considered unacceptable (and a supposed consequence of apartheid).
Although the differentials have grown enormously since then, employers are still required to provide this information every year, using a prescribed form. One may ask why the government keeps this information secret, for the data reveal the management income that is elsewhere concealed as wages. But we should also note how ignoring this data helps economists (mis)use average so-called wages.
Statistics South Africa’s “Monthly Earnings of South Africans 2010” provides another rich source of highly relevant data. It gives comprehensive details on median wages as opposed to the misleading average wages. The “median” is the appropriate measure of anything that is highly spread out, such as South Africa’s notorious inequality between the very rich and a very poor. Median calculations addresses this distorted picture, shifting an exclusive focus on wages to the counting of people, with wages being the discriminator.
The median wage is thus the wage that distinguishes the bottom 50% of the population from the top half. If, in the previous example of 100 people, 50 earned R5 or less and the remaining 50% earned between R5.01 and R1000, the median wage would be R5. This is why the median is also called the “typical wage”.
According to Statistics South Africa’s monthly earnings survey, the median wage for a South African worker, formal or informal, was R2800 a month in 2010. That is, half of the total of a little more than 11-million ordinary employees in South African earned R2800 or less.
Making it even easier for economists to give accurate information, the monthly earnings report seeks to give data about people who are more typically seen to be workers. It thus excludes two categories usually classed as workers. It excludes the 686000 people who employ others. Their median earnings are R7000, which is so low it indicates that the owners of the large number of small and medium enterprises are covered by this group. Also excluded are the 1207000 self-employed or “own-account workers”. They range from the corner car guard to the previously mentioned solicitors who earn R40 000 a day. (Despite the wealth at the top end of this group, the earnings median is a mere R1820 a month.)
Half the informally employed workers earned R1600 a month or less. Half the 568000 workers in agriculture earned R1213 or less and half the domestic workers – about 565 000 people – earned R1000 a month or less.
What this outline shows is that there is no excuse for our celebrity economists to misuse statistics. They ignore available data. They rely on the public’s general discomfort with statistics to get away with selling their prejudices.
The question is whether this is done knowingly or not – and which is the more reprehensible.
If, or to the extent that it is not done knowingly, these economists could be well served by exposure to Statistics for Dummies.
Dr Jeff Rudin and Dr Dick Forslund are researchers at the Alternative Information and Development Centre in Cape Town