Print Page
News & Press: Opinion

High income-tax rates drive firms into informal sector

Monday, 25 February 2013   (0 Comments)
Posted by: SAIT Technical
Share |

By Loane Sharp (Business Day)

Executive summary

This article examines discrepancies between the number of taxpayers estimated by Stats SA and the actual numbers reported by SARS. The author believes that people appear to be underreporting their incomes in Stats SA surveys. If the tax burden for high-income taxpayers goes up, the economic incentives to disappear into the informal sector will go up with it.

Full article

SOUTH Africans have been led to believe by Statistics South Africa that the country’s informal sector consists mainly of retail activity at traffic lights and in spaza shops. According to Stats SA, employment in the informal sector consists of employees who work for a small business (employing less than five people) and who are not registered for income tax (they earn less than R60,000 a year). Of the estimated 2.2-million workers in the informal sector, ostensibly 46% are employed in retail trade, 15% in personal services and 14% in construction. Stats SA’s data are almost certainly wrong.

Finance Minister Pravin Gordhan indicated in last year’s budget speech that there are 6.2-million payers of income tax in South Africa. Income tax is theoretically assessed on 11-million people, but 4.9-million fall below the tax threshold, with the result that income taxes are paid only by 6.2-million people.

This is strange as, according to Stats SA, there are 13.1-million employees, employers and own-account workers in South Africa. What accounts for the 2.1-million "missing” taxpayers? Or, stated differently, why are 16% of income earners not paying income tax?

It is notable that the discrepancy between the estimate (from Stats SA) and the actual number (from the South African Revenue Service) of income-taxpayers rises proportionally with reported income. For example, 102,000 people in the income band above R1m pay income tax (according to SARS), yet there are only 35,000 people in this band (according to Stats SA) — a discrepancy of 67,000 people. By comparison, 1.8-million people in the income band from R60,001 to R160,000 pay income tax (according to SARS), yet there are as many as 2.8-million people in this band (according to Stats SA) — a discrepancy of 1-million.

Stats SA does not make proper use of so-called administrative data (nonstatistical data, such as the number of income-taxpayers reported by SARS) to adjust the estimates derived from its sample-based surveys. No serious statistics agency would tolerate the sort of discrepancies that exist between Stats SA’s data and the administrative data collected from SARS, the Council for Medical Schemes, the Unemployment Insurance Fund, The Business Trust and others.

The connection to the informal sector is this: people appear to be underreporting their incomes in Stats SA surveys. As a whole, 4.5-million people report to Stats SA enumerators that they earn income in the taxable range, whereas 6.2-million are actually paying income tax to SARS — a 27% undercount by Stats SA. Or, put differently, the average taxpayer’s income is R266,641 a year, according to SARS, and R193,325 a year according to Stats SA — a 28% undercount.

Without access to additional data, it is impossible to know to what extent individuals underreport their incomes to SARS. However, we can come up with a good estimate. For one thing, we know tax evasion occurs. A recent study by University of Stellenbosch economist Philip Black suggests that, following the increase of taxes on cigarettes from 12c to 45c a cigarette between 1999 and last year, smuggled cigarettes now constitute 45% of total cigarette consumption.

SARS and the National Treasury declined to give us data on fuel taxes, but it is likely that, after the escalation of fuel taxes on petrol to more than 30% of the official pump price, the proportion of smuggled petrol in South Africa is now 10%-20%.

Additionally, we know there is a poor correlation (just 25%) between national income and income-tax revenues. This implies that the income tax regime in South Africa is complex, as changes in income explain only a small proportion of changes in the value of income taxes collected. We also know that income taxes have grown substantially relative to incomes, which greatly increases the tax burden: 1.7% pay 24.3% of total income taxes, and 4.5% pay nearly 40%. Both these factors — the complexity of the income tax regime and the disproportionate incidence of income tax on people earning more than R600,000 a year — give rise to plenty of economic incentives to evade taxes.

One available procedure involves relating income and wealth to the taxes collected on income and wealth. If the income/wealth of all individuals is known by the tax collector with perfect certainty, and if income/wealth taxes are collected perfectly efficiently, there should be a simple relationship between them — the average tax rate. In order to perform this computation, we collected data on income and wealth taxes (of households, for example), which is straightforward, as the Reserve Bank supplies such data. Also, we developed an indirect measure of income and wealth as wealth, especially, is a highly subjective concept, including as it does hard-to-measure human capital wealth. Following University of Cape Town economist Brian Kantor, we used household consumption spending as a proxy for household income and wealth, and we applied a smoothing procedure to tax collections, as taxes on income/wealth are very volatile.

There is nowhere near a close relationship between household income/wealth and taxes on household income/wealth. The correlation between the two is 71.2%, and this only because we have smoothed the tax series considerably. Without smoothing, the correlation is just 12.9%. Moreover, the correlation has not changed to any appreciable degree between the (apartheid-era) Commission for Inland Revenue (67.4%) and the (post-apartheid) SARS (69.3%).

We draw two inferences from this. First, there is tremendous inconsistency in income-tax collection. Using the level of reported income for the economy as a whole, the authorities know only vaguely what the level of household income is, and to the extent that they know it at all, they do so only on average and with three to five years of hindsight.

Second, the authorities are collecting only two-thirds to three-quarters of the income taxes owing to them. Households have succeeded in hiding between a quarter and a third of their incomes, a figure that has not changed appreciably for the past 50 years.

The implications for our understanding of the informal sector are significant. According to Stats SA’s Quarterly Labour Force Survey, 440,000 small-business "closures” occurred between 2006 and 2011, and the number of new business startups fell to a record low last year. We put the term "closures” in inverted commas, because it is likely that these businesses continued to exist and operate but ceased reporting their activities to the authorities. These enterprises probably re-emerged in the informal sector, where compliance with income-tax laws is patchy. It is no coincidence that SARS’s tax amnesty for small businesses and sole proprietors between 2003 and 2010 netted about 385,000 additional income taxpayers.

The lesson we draw is that South Africa’s informal sector consists not only or even primarily of hawkers, but of a growing number of moderately high-income individuals who, owning their own businesses, have many options to evade income taxes.

In his recent state of the nation address, President Jacob Zuma warned tax rates will rise in the coming years. If, as seems probable, the tax burden for high-income taxpayers goes up, the economic incentives to disappear into the informal sector will go up with it. South Africa’s informal sector is growing, largely due to circumvention of labour laws and evasion of income taxes. In order to stem the growth of informal economic activity, there is no real alternative but to relax labour laws and reduce income taxes.

Sharp is a labour economist at Adcorp.



Section 240A of the Tax Administration Act, 2011 (as amended) requires that all tax practitioners register with a recognized controlling body before 1 July 2013. It is a criminal offense to not register with both a recognized controlling body and SARS.

  • Tax Practitioner Registration Requirements & FAQ's
  • Rate Our Service

    Membership Management Software Powered by YourMembership  ::  Legal