LOANE Sharp, economist at Adcorp, the recruitment company, on Friday upped the ante in his continuing argument with Statistics SA and two academics over how employment is measured in South Africa.

The government uses Statistics SA's figures to, among others, justify the continual broadening of the South African welfare state.

More than 15-million of an estimated 50-million people in Africa's biggest economy receive some kind of state welfare.

Mr Sharp on Friday presented evidence that Adcorp's statistics present a more accurate impression of prevailing economic conditions.

Stats SA puts unemployment at 24,9%. Adcorp says only 9% of South Africans are not economically active or earning a regular income.

Mr Sharp said Bank of America Merrill Lynch analysts, who used careful analysis of aerial photographic records, suggested the size of the economy was "possibly substantially underestimated".

"Actual VAT (value-added tax) payments to the South African Revenue Service (SARS), compared with hypothetical VAT payments, calculated by the National Treasury, suggest that the economy's size, notably household consumption, accounting for 64% of gross domestic product, is substantially underestimated."

Mr Sharp said FinScope, a survey conducted under the auspices of The Business Trust, estimated there were 5,8-million more small business owners or employees than recorded by Stats South Africa.

Statistics SA last year questioned Adcorp's statistical methods.

Recently, the company was criticised by two University of Cape Town academics, Martin Wittenberg and Andrew Kerr, who run the university's DataFirst survey research unit. They said Adcorp used crude estimates to calculate unrecorded activity and then guess the level of labour intensity needed to produce that level of activity.

Stats SA uses a periodic questionnaire and estimates informal sector employment is 2,1-million.

According to Mr Sharp, the most compelling evidence against Stats SA comes from its own data.

He said last year 3,6-million people failed to declare to Stats SA field workers that they received government social grants - an undercount of 25%. According to Mr Sharp, between 2003 and 2010, SARS figures showed the number of personal taxpayers rose by 2,5-million people, or 74%, and the amnesty programme for small businesses netted 355000 new business taxpayers, which was an undercount of about 18%.

Renaissance Capital economist Elna Moolman says investors trust Stats SA data as it complies with internationally accepted standards. Investors take jobs data seriously.

"For an investor, the absolute level is important from a structural perspective and it has implications for the fiscus and sociopolitical stability, among other factors. So at very high levels of unemployment and poverty, all other things equal, the risk that there might ultimately be some sociopolitical uprising against this is higher.

"It also places a drain on the fiscus and narrows the tax base, implying higher taxes on the tax-paying population, which affects spending power and/or profits," she says.

In terms of tactical investment decisions, the changes in employment were more important.