SAGA Progress Report - South Africa

SAGA
B16 MVR Hall
Ithaca, NY 14853
(607) 255-8931
Fax (607) 255-0178
saga@cornell.edu

SAGA Progress Report
October, 2002

II. RESEARCH
5. South Africa

We propose specifically, to organize the research program and its various sub-elements under five major themes.

A. The Evolution of Poverty and Inequality: A Base Analysis

The purpose of this theme is to provide an empirical assessment of the altering nature of poverty and inequality in the post-apartheid South Africa. Some of the broader issues within this theme, which is essentially a measurement of the altering poverty and inequality dynamics over the post-apartheid period, would entail the following:

Are the standard South African markers for poverty and inequality (race, gender, location) increasing or declining as significant correlates of poverty and inequality?
Does the inequality decomposition analysis reveal the same pattern in terms of sources of incomes and their contribution to inequality?
The correlates of labor market vulnerability on the 1995 data are now well established. The time comparison here would allow one to test for any changes that may or may not have occurred in the prevalence of vulnerability within the labor market.
One of the key advantages of the surveys is the sets of questions on access to services, such as health, education, sanitation and housing. Significant improvements in the access to services (controlling for quality, cost and so on) would be vital for a broader understanding of successes or failures in poverty alleviation.
The post-apartheid period has seen an expansion in the national transfer income program of government, marked by steady increases in the value of the old age pension. One of the key considerations is that, as one of the most effective mechanisms for poverty alleviation in the society, has the overall income grant system had a significant poverty reduction effect, controlling for other sources of income?

B. The Economics of Education and Education Policy

We will concentrate our research around two broad issues:

The determinants of success in both the schooling and educational system.
The nature and level of distributional equity (measured in a variety of different ways) in higher education and schooling and how this impacts on success.

In terms of the first of these, we would identify firstly a measure of success in both the schooling and higher educational system. In the first instance, the integrated databases would allow for a presentation of the various descriptive statistics on success rates, according to a variety of different markers including - for schools - the province that the school is located in, the dominant racial group in attendance, pupil-teacher ratios, per capita expenditure, and so on. The multivariate analysis would then try and determine, across all nine provinces, the importance of these different covariates in determining pass rates at schools.

In terms of higher educational institutions, the tracer study database would present information on the institution attended by students and the type of employment they found. The univariate analysis would isolate the success rates of various HED institutions in the sample, while the multivariate would try and combine both the supply characteristics of the individual, with the characteristics of the institution in trying to determine probabilities of success.

The notion of equity goes to the heart of the post-apartheid debate on education policy, where the focus has been on trying to alter past unfair distribution of resources allocated to schools, universities and technical schools, as a key mechanism for altering inter-institutional inequities. Firstly, we propose to measure the level of distributional equity in the system, according to variety of markers including, per capita expenditure, pupil-teacher ratios, user fees, teacher qualifications, direct assets, and complementary assets. Each of the markers would be derived according to the nature of the school or HED institution (identified as historically white, African, urban or rural, within a high income area for schools and so on). This procedure ultimately delivers a matrix of resource distribution across the entire education system in the country.

C. Health and Nutrition

The analysis of this theme would examine the access to health care, utilization of health services, and health status in relation to household poverty. Coupled to this, would be an analysis of the impact of the HIV epidemic on households.

On the first of these, given the demographic and epidemiological transition in South Africa, diseases related to poverty such as tuberculosis, child mortality and diseases related to rapid urbanization such as injury-related deaths, chronic diseases of lifestyle such as obesity, hypertension, and diabetes are on the rise. There is, however, a dearth in information that links these health status measures with socio-economic variables such as household income, access to services, and education levels. Hence, the first element of this theme would examine poverty in relation to a set of key health status indicators through univariate and multivariate analysis.

This theme would examine the relationship between access to and utilization of health services according to the standard range of socio-economic covariates, including race, gender, location, education levels, and income levels. Specific variables would be examined such as the type of service attended (private/public sector), access to private medical aids, the reasons for seeking care, and from whom care was sought. Where possible, client satisfaction and perceptions will be explored.

Finally, the HIV epidemic remains the most serious public health problem in South Africa. It is well established that the epidemic affects young, heterosexual adults. However, there is there is a dearth of robust information on the social and welfare correlates of the pandemic. Furthermore, the link between HIV/AIDS prevalence and the various determinants of poverty has not, thus far, been explored adequately in the South African literature. There is little information on how the distribution of HIV disease would affect employment, income distribution, savings rates, consumption patterns, poverty, and other economic variables. The purpose of this theme would be, through the use of household survey data and the HIV Antenatal Survey data sets, to provide an overview of the features of the HIV/AIDS epidemic at the household level. Specific factors such as poverty levels, education, dependency rates and employment levels would be examined in relation to HIVAIDS.

D. Risk and Vulnerability

This theme is potentially enormous, given that it lies at the nub of much of the work on poverty traps and analyses on the dynamics of poverty. We propose to focus here on three key sub-themes that we view as pertinent for South Africa:

Price shocks and their impact on poor households
Crime, poverty, and inequality
Mobility within the internal labour market and between employment and unemployment

On the first theme, there is evidence to suggest that the construction of the national consumer price index is heavily weighted in favor of the upper deciles of the income distribution. We do not therefore have a good sense, firstly of the consumption bundles of the poor and additionally, how price movements and shocks impact on their livelihoods. An examination of the IES95 and IES00 therefore, would be a first attempt at constructing these consumption patterns across the distribution, and, of course, to examine changing patterns of consumption. When matched with the disaggregated price data, we would derive estimates of the inflation rate for all deciles, focusing on poor households. The two points of price and expenditure data, would thus also allow for estimation of price elasticities of demand across the different deciles and product categories.

In terms of the second sub-theme, we would utilize the two national databases of Statistics South Africa on crime, as the basis for research here. What we will attempt to assess here is to determine the extent to which poor households are affected by crime, and the extent to which this activity serves as an unforeseen shock to household incomes and assets. While the current work on crime, inequality, and poverty is limited in South Africa, we would hope that this sub-theme could gradually form the basis for expanded work in the area and indeed the manipulation of new data sets.

One of the key elements engendering vulnerability amongst individuals and the households they reside in, is the uncertainty in employment status and earnings derived from employment. This labour market instability therefore, can be viewed as a key determinant of household risk and vulnerability over time. South Africa, since 2000, has been running a bi-annual Labour Force Survey (LFS), which has a rotating panel as a sub-sample. Currently, the LFS is in its 6th wave, and early indications are that the data is at least being released timeously. While its panel component would still need to be verified and tested, it is wholly possible that the LFS offers the advantage of undertaking dynamic panel analysis of labour market risk and vulnerability. As a starting point, for example, it would be possible to examine through a standard mobility matrix, the movement of individuals across the wage distribution, into and out of employment and across the household income distribution.

E. Labor Markets and Poverty

We view this theme has having at least five sub-components, namely:

Labor market-household dynamics in the post-apartheid period
Education and labor market dynamics in the post-apartheid period
Labor demand shifts over time
The unemployed, the unemployable and household poverty traps
Poverty alleviation and job creation strategies

Institutional arrangements

The SISERA partner institution, the Development Policy Research Unit (DPRU) is a research unit recognized by the University Research Committee. In addition, the unit has a very strong record in policy-orientated, yet academically sound, research.

Activities over the previous six months

After the selection of South Africa as a core country, discussions began with the SISERA partner institution DPRU in June, 2002. In keeping with the demand driven philosophy of SAGA, discussions focused on DPRU’s research priorities, in light of the policy needs in South Africa, with input from the USAID mission.

The policy conjuncture in South Africa is different now from even a year ago. As confirmed by the USAID mission and DPRU, there are growing concerns in the country and in the government that the macroeconomic reforms and opening up are not delivering, or not delivering fast enough, on poverty reduction. The very high rates of unemployment are a particular concern. This conjuncture comes at a time when the government is putting together its 10 year perspective strategy, and at a time when there is discontent among the ANC’s governing partners about the state of poverty.

These concerns coincide with the imminent general release of the 2000 living standard survey, a five-year follow up to the 1995 survey, which will give a snapshot of the first five years of post-apartheid government. In discussions, it emerged that DPRU’s research plan for the next two or three years envisages a "post-Apartheid audit" anchored on these two major data sets, leading to detailed research on a number of key dimensions of poverty.

It should be emphasized that this is DPRU’s research plan, and funding for it, will come from a number of different sources. DPRU will apply to SAGA-SISERA's competition to fund research on Poverty 1995-2000, and to the African Economic Research Consortium (AERC) to fund research on Labor Markets in South Africa. SAGA-Cornell will support the overall research strategy as well as providing initial seed financing for it. It is envisaged that SAGA-Cornell will focus its major funding on the last three themes—on vulnerability, education and health—as the research program gets more sharply defined.

Activities anticipated over the next six months

DPRU will apply to SAGA-SISERA and to AERC. SAGA-Cornell will help with these proposals. DPRU and Cornell will sign the first contract for specific use of SAGA-Cornell funds; some of this will be seed money to start work on analysis of the 1995-2000, and some of it will be funding to begin work on one of the SAGA related themes—at this stage it looks as though the initial focus will be on the risk and vulnerability theme. DPRU will also begin preparing for a training workshop for South African researchers on poverty analysis, for which they will access funds from SAGA-SISERA. SAGA-Cornell technical assistance funds will also support the preparation of this activity.

Previous Section | Next Section

Return to SAGA Progress Report (October 2002) Table of Contents