I think it’s great that this debate is raging.
Briefing note released by Policy Co-ordination and Advisory Services – Global Insight, South African Institute of Race Relations (SAIRR) and Poverty: Whose Reality?
22 November 2007
The South African Institute of Race Relations (SAIRR), citing Global Insight Southern Africa (2006) reports that “[t]he number of people living on less than US$1 per day (the measure of extreme poverty used by the World Bank) in South Africa, increased from 1,89 million in 1996 to 4,2 million in 2005. As a proportion of the population, this represents an increase from 4.5% to 8.8% of the population living on less that US$1”.
This report differs dramatically from reports by various distinguished academics in the field.
From the onset, it is important to note that officially there is no consistent time series data on poverty and inequality trends over time. Most researchers use a combination of Statistics South Africa official data and unofficial data sources whose credibility they can guarantee, e.g. All Media Products Survey (AMPS) to arrive at a view on poverty levels and trends. They also adopt different poverty lines, ranging from about R2000 to R4000 per annum. Global Insight probably also uses a combination of data sources. However, there is no reflection on their sources and methodologies in the report.
What then should guide our understanding of the realities of poverty in South Africa? Perhaps the starting point should be the South African Constitution and the socio-economic rights it asserts should be progressively realised. These include access to housing/shelter, education, social assistance, health provision, food and water. This starting point, by implication, acknowledges that taking a narrow view of poverty and using a simplistic measure such as money-metric poverty barely engages with the reality of poverty and thus the combination of things that should be done to deal with it. The Constitution adopts a multi-dimensional approach to poverty that incorporates the meeting of basic needs.
Money-metric poverty
Let us begin with money metric poverty, the same indicator used in the SAIRR survey. How has this evolved over this reported period and beyond? Different reports are used here to illustrate the inconsistent view of the SAIRR report with the views of major work done in this area. Amongst this is work by UNDP covering the period 1995 to 2002. Using a national poverty line of R354 per month, the UNDP reported that the proportion of people living in poverty fell from 51.1% to 48.5% between 1995 and 2000. However, the absolute number of people living in poverty increased from 20.2 million to 21.9 million as a result of population growth. Other work by Hoogeveen and Ozler (2004) covering the same period, reports stability in the proportion of people in poverty using the R322 poverty line, and an increase in the absolute number of people living in poverty due to population growth.
Using a poverty line of between R3000 per capita per annum (in 2000 constant Rand) Van der Berg et al. report a slight increase in the proportion of people living in poverty between 1993 and 2000. However, this declined substantially even in terms of absolute numbers – from 18.5 million in 2000 to 15.4 million in 2004. Extending the period further to 2006, as is shown in the table below, headcount poverty significantly decreased from 2001 to 2006. In the entire period, 1993 to 2006, there was a proportionate decrease from 50.1% to 43.2%.
PERCENTAGE OF POPULATION LIVING BELOW R3 000 PER ANNUM
Year & Poverty headcount index 1996: 53.1% 1997: 51.2% 1998: 51.0% 1999: 52.1% 2000:50.8% 2001: 51.4% 2002: 49.0% 2003: 47.6% 2004: 46.9% 2005: 44.5% 2006: 43.2%
Van der Berg et al. also reflect on the changes in the poverty gap and severity of poverty. They show that the poverty gap index declined from 0.26 to 0.21 during the period under review (1996-2005). This implies a significant improvement in income and expenditure of those in poverty, bringing them closer to the poverty line of R3000. Furthermore they report a reduction in the depth and severity of poverty, from 0.16 to 0.12 in the same period. Calculated as the square of the difference between the poverty line and the incomes of the poor, the measure of depth and severity of poverty gives greater weight to those who are most deeply in poverty; and the fact of its improvement is even more significant.
Furthermore there is evidence that income poverty levels are indeed declining in part because of improvements in employment levels, but more importantly because of the enormously improved reach of the social grant system. Between 2000 and 2004 the per capita income of the poorest two population quintiles increased by over 30%.
Reaching the bottom 40% of the income distribution, government expenditure on cash transfers in the form of social grants increased from R10 billion in 1994 to R55 billion in 2005 and the number of beneficiaries grew from 2.6 million to 10.5 million. With the extension of the Child Support Grant to children under the age of 14, the total number of children accessing the grants is now over 7 million.
The impact of improvements in the labour market on poverty cannot be overstated. The number of employed people increased from 11.1 million in 2001 to 12.8 million in 2006.
The social wage
Beyond income poverty, which relates to private income and cash transfers, is the social wage. It will be folly to ignore the analysis of this component in the monitoring of progress or lack thereof with regard to poverty in our country. Emphasising the importance of the social wage, the HSRC made the following recommendation:
“The magnitude of the social wage is so great relative to poor people’s incomes, and even to the grants they receive, that any ongoing appraisal of government performance in supporting the poor, must take it into account…. [T]here would be value in ensuring the ongoing tracking of the social wage, with further methodological refinement.” (HSRC, 2004, p.32).
What are some of the elements of the social wage and how has it changed over time? The recent Community Survey provides a pointer in this regard. It finds that the proportion of households who use electricity increased from 56% in 1996 to 80% in 2007. The proportion of households who had access to piped water in dwelling or on site increased from 61% in 1996 to 70% in 2007. Households with access to flush toilet has increased from 52% in 1996 to 60,4% in 2007.
Another measure of the social wage is the government’s role in improving access to assets such as housing and land. According the Community Survey, the number of households living in formal dwellings increased from 69% in 1996 to 71% in 2007. More than R50 billion of assets (in the form of subsidised housing and land) was transferred to poor households in the period 1994 to 2003, according to the government’s Ten Year Review.
How well targeted are these programmes? Bhorat et al., although reporting on an earlier period, provide a sense of the beneficiaries of these improvements in services by income deciles. They indicate that during the period 1993 and 2004 access to formal housing grew by 42% and 34% for income deciles 1 and 2 respectively (the poorest sections of society) and 21% and 16% for deciles 3 and 4. Access to piped water increased by 187% in decile 1, while the growth was 31% in the 4th decile; and access to electricity for lighting for those in decile 1 grew by a phenomenal 578%. As the government’s programme of action indicates, 74% of those with access to water also receive free basic water, and the figure stands at 59% for electricity.
Account should also be taken of the education and health programmes, which include free medical care for pregnant women, children under six and people with disability; free text-books, improved infrastructure and no-fee paying schools
Using 2003 data (and Rand value), an HSRC study estimated that about 50% of this gross social wage value was directed at households in the poorest 40% of the population. This suggests that, taken as a whole, the programmes are progressively targeted. For social grants, 62% of the total went to the poorest 40% of households, and 82% to the poorest 60% of households. Furthermore, the HSRC study reports that female-headed households, those comprising single women supporting children, and granny households, receive larger-than-average social wage totals.
The benefit incidence demonstrates the impact of government programmes aggressively to address poverty and inequalities which reflect a history of systematic discrimination. This pro-poor bias of government social expenditure also demonstrates the critical role of the fiscus as an instrument of redistribution of wealth.
Exchange rate and purchasing power
Perhaps the central flaw in the SAIRR (Global Insight) report is the use of $1 per day as a poverty line. While this poverty line is recommended by the World Bank and used by most international agencies such as the UN, the uncritical adoption of this by SAIRR is unfortunate. This is particularly so in the context of the problematic nature of this measure especially with regard to exchange rate translations. Further, it is highly complex to correct for the differences in the relative purchasing power of the Rand to the US dollar.
But irrespective of these complexities, the report that the proportion of those below this poverty line has increased in the period under investigation is in contrast to evidence presented in other work. The All Media and Products Survey data for example, shows that the poorest income group (Living Standard Measure, LSM 1) fell from 10.5% to 4.8% of the population between 2001 and 2006. The bottom three income groups (LSM 1-3) shrank from 38.8% to 27.7% of the population in this period. Furthermore the incomes of people in the poorest groups grew by about 38% between 1993 and 2004, mostly as a result of the expansion of social grants.
The inconsistency of the SAIRR report calls into question their selection of particular data sources and their processing thereof. On a matter as critical as poverty-reduction, one would have expected at least a rigorous process of peer review and wider consultations before publication. In the event, the SAIRR in its carelessness and haste has served to generate more heat than light.