Abstract
This article provides statistical estimates of the level of relative poverty over time in the Western Cape province of South Africa, using data sets from Statistics South Africa. It seems that relatively large numbers of Black and Coloured households are chronically poor, and that substandard education and living conditions are likely to be causing this situation. The authors propose short-term and long-term poverty alleviation plans that aim to increase accessibility to social services and to raise educational levels, to ensure sustainable livelihoods for the people concerned.
1. Introduction
Poverty is generally seen as a major characteristic of underdevelopment, and poverty elimination is the main purpose of economic policy in most developing countries (Deaton, Citation1998). President Mbeki pledged his government to create jobs and fight poverty in his Year 2001: State of the Nation address: ‘The levels of poverty, unemployment and underdevelopment in some parts of our country are too high’.Footnote2 Job summits were convened in 1998 and 1999 at national and provincial levels and provincial governments committed themselves to formulating new priorities in budgetary expenditure so as to maximise job creation. Poverty is multidimensional in nature; the poverty problem should therefore be addressed primarily through satisfying basic human needs within an integrated and broader multidimensional programme of sustainable development that ensures delivery (see NRF, Citation2000).
In South Africa and particularly in the Western Cape province, which is the subject of this paper, relatively large numbers of people are absolutely poor, and the distribution of income is very unequal. The Gini coefficient is 0,59, which is very high by international standards (Deaton, Citation1998). The incidence of income inequality is most pronounced between race groups, with the average income of Whites estimated to be almost 11 times higher than that of Blacks (Whiteford et al., Citation1995: 14). Using the 1991 Census data, Whiteford et al. Citation(1995) found a Gini coefficient of 0,68 for South Africa.
Labour market discrimination during the apartheid years contributed greatly to this state of affairs (Bhorat et al., Citation1995: 1). Although inequality and discrimination between the races declined during the 1980s with the abolition of apartheid, inequality rose within race groups (Knight & McGrath, Citation1987; Moll, Citation1995). Corroborating earlier evidence by Knight & McGrath Citation(1987), Serumaga-Zake & Naudé Citation(2001) found that labour market discrimination against Blacks is not statistically significant, and that the productivity of Whites explains a large proportion of the income differential between races. This suggests that the effects of previous and persistent pre-labour market discrimination, in particular differential access to education and health facilities, are the most significant impediments to reducing income inequality. In 1993, the poverty head-count ratio in South Africa was 0,317 for Blacks and 0,250 for all races, using a poverty line of R105 (or US$10) per person per month (Deaton, Citation1998).
There is also a significant spatial or regional dimension to income inequality in South Africa (WEFA, Citation1996: 41–5). The province with the lowest Gross Geographical Product (GGP) per capita in 1995 was Limpopo (formerly the Northern Province) (R2 721), and that with the highest GGP per capita was Gauteng (R23 215). The average income in the richest province (Gauteng) is almost 10 times that of the poorest province (Limpopo). Moreover, indicators of access to education and health such as the Human Development Index are almost twice as high in Gauteng and the Western Cape as in Limpopo (Whiteford et al., Citation1995). The unemployment rate also differs by province. Employment opportunities are relatively abundant in the Western Cape and Gauteng, the provinces with the highest Human Development Index scores.
As elsewhere in the world, unemployment is higher among females than among males. According to WEFA (Citation1996: 41–5), the unemployment rate for females is more than 15 per cent higher than that for males in South Africa. This is caused in part by economic discrimination against women and by cultural norms and values (Serumaga-Zake & Naudé, Citation2001).
According to Naudé & Serumaga-Zake Citation(2001), provinces such as the North-West province are characterised by relatively low participation rates and high unemployment rates. These in turn depend on residence in rural or urban areas, location within a region, age, gender, race, education and the type of employment which can be found. Serumaga-Zake & Naudé Citation(2002) found that education and household size were significant determinants of poverty in the North-West province, and that different determinants of welfare existed between the poor and non-poor.
1.1 Government policies vis-à-vis poverty alleviation
It is of great concern that the current policies of the national government hold the potential to exacerbate spatial inequality in employment opportunities. For example, the national government embarked on a programme of substantial tariff reduction. Coetzee et al. Citation(1997) concluded that tariff reduction would be likely to hamper the country's poverty alleviation programme by reducing employment and output in industries such as textiles, clothing, food processing and metal processing that are relatively more widespread in the poorer provinces of Limpopo and the Eastern Cape. High labour intensity, relatively unskilled labour and high rates of both nominal and effective tariff protection in former years characterise these sectors. Johnson Citation(1997) argues that trade liberalisation can increase income inequality, since changes in relative prices caused by liberalisation will shift domestic production of tradable goods away from unskilled-intensive goods towards skill-intensive goods. This forces unskilled labour to crowd into the non-tradeable sector and causes a decline in their relative wages [see also Freeman Citation(1995), Richardson Citation(1995) and Wood Citation(1995)].
In terms of Schedule 4 of the Constitution of the Republic of South Africa (adopted 3 February 1997), the provincial governments are responsible for the provision of education, health and welfare services, all of which may lower income inequality. However, the provinces have little power to tax. Moreover, the provinces were forced to reduce their expenditure on education and health to achieve the national government's aim of reducing the budget deficit from its level of 4,1 per cent of Gross Domestic Product (GDP) in 1997/98 to 3,0 per cent in 1999/2000.
In South Africa, poverty varies from province to province because of large differences in economic structure between the provinces. Provincial governments need to be able to formulate their own economic development strategies in light of their particular budgetary constraints. These strategies should be based on reliable statistical data and estimates of the determinants of labour force participation, unemployment and poverty.
In this article, poverty is considered as relative material deprivation and therefore will be measured in terms of satisfying the basic human needs and accessibility of social services such clean water, energy, education and health. The government of South Africa has for some time put in place socio-economic policies and programmes to alleviate poverty, such as affirmative action, feeding schemes for school children, free medication for pregnant women and children and programmes to promote education and skill acquisition. This study was intended to investigate the dynamics of relative poverty in the Western Cape province. The study specifically seeks to uncover who the poor are and to distinguish between transient and chronic poverty in the Western Cape province.
The results suggest that relatively large numbers of Black and Coloured households are chronically poor, and that poor education and living conditions are responsible. Households headed by Black women are the most vulnerable.
The outline of the paper is as follows. Section 2 explains the methodology. The results are presented in section 3. Three sub-sections deal in turn with basic human needs, the possible causes of poverty and the distribution of monthly total household expenditure (as an indicator of wellbeing) in the Western Cape province. Conclusions and suggestions for future research are given in section 4.
2. Methodology
This study used data sets for the years 1995 to 1999, which were generated from the October Household Surveys (OHS) on socio-economic characteristics, which are conducted by Statistics South Africa. Footnote3 These were the only data available to the authors, and all the conclusions in this paper are based on them. The authors had nothing to do with the development of the survey forms, the method of data collection, the actual data collection, or the construction of the data files.
Combining data from the OHS for the years 1995 to 1999 is a difficult task. There are several reasons for this. Firstly, the questions asked in the surveys changed from year to year. The wording of the questions and the location of the questions within the survey changed. The variable name assigned to the questions also changed. For example, the variable for race was named D_RACE in 1996, 1998 and 1999, PRACE1 in 1997, and Q2_1POPG in 1995. Other variables had different names in each of the five years. In combining the data from the different years into a single data set it is necessary to have a common name for each year.
A second difficulty with the data is with changing response options. These changes required extensive recoding of the data. For some variables, the set of options remained the same, but the numeric codes differed. For example the code for relationship to head of household being a brother or sister was 7 in 1995, but 4 in other years. For other variables the set of options changed. For race, there were four options in 1995 and 1996 (African/Black, Coloured, Indian/Asian and White).Footnote4 In 1997 and 1998 two additional options were added (Griqua and Other), while in 1999 the option of Griqua was removed. For other variables the options changed in more complex ways, so that judgement had to be used in combining categories so that they were roughly equivalent across years. Variables whose categories could not logically be combined as equivalent were not considered for use in this study. It is possible that some ways of combining categories that appeared ‘logical’ were in fact not so, resulting in variables that are not truly comparable across years. This is recognised as a limitation of this analysis.
Other difficulties in combining the data across years relate to missing values. In 1995 a missing value was often coded as 0 while in 1996 a code of 0 was also frequently used, although for some variables a code of 5 was used. In 1997 we observe that 9 and 99 were also sometimes used for missing values. In the 1998 and 1999 data records, @ was used to indicate ‘not applicable’ and * was used to indicate ‘missing’.
After resolving issues of variable names, coding differences, and missing values across years, the data were analysed using the Statistical Analysis System (SAS) software. The surveymeans Procedure was the primary procedure used for the analysis. The weights used were those given for the household as the analysis was conducted at the level of the household. In the results and tables, point estimates of the proportions of people or households having given characteristics are given. We have chosen not to give confidence interval estimates of the proportions due to space limitations. Due to the relatively large sample sizes in any year, the confidence intervals are not very wide. Typically for proportions near 0,5 the upper and lower limits of the population proportions would be found by taking the estimate plus or minus a value ranging from 0,05 to 0,21.
3. ResultsFootnote5
3.1 Type of dwelling, sanitation, clean water and energy over time
3.1.1 Type of dwelling
The type of dwelling was classified into one of three categories: brick/semi-detached, shack/huts and flat/other. indicates a rising trend for the percentage of Black households living in shacks or huts in urban informal settlements (colloquially known as ‘squatter camps’). Overall (see ), the percentage of urban households living in shacks or huts rose from 10 per cent in 1995 to 13,5 per cent in 1999.
Figure 1 Percentage of Black and Coloured households living in huts or shacks in urban areas of the Western Cape province, 1995–99 (per cent)
Table 1 Type of dwelling occupied by urban and rural households in the Western Cape province, by race, 1995–1999 (percentage distribution)
After the new democratic government came to power in 1994, Blacks migrated in large numbers, especially to areas in which they had previously been denied the right to live and to areas from which they had been forcibly removed during apartheid. It is common knowledge that many large informal settlements were established and grew dramatically in or near the cities during the 1990s. The scatter diagram () shows a large jump in the percentage of Black households living in informal settlements in urban areas between 1995 and 1996, and a decline after 1996. This subsequent decline may reflect the efforts of the government to contain the difficult situation with its massive housing scheme. If the 1995 point observation were excluded, the rate would show a consistent decline. The housing situation for Coloureds in urban areas is also alarming, with a constant 4 per cent living in shacks or huts over the period.
In rural areas, the situation was not equally bad; the number of households living in shacks reached a peak of 16,4 per cent in 1996 and fell to 2,8 percent in 1999.
3.1.2 Sanitation
The type of sanitation was classified into one of three categories: in dwelling, on-site and off-site. shows that urban households with off-site sanitation increased from 1,8 per cent of the total in 1995 to 8,4 per cent in 1999. Of the urban households with off-site sanitation about 83 per cent were Blacks and 13 per cent Coloureds. The proportion of households with off-site sanitation increased over the period from 1 per cent to 17,4 per cent in rural areas. Here Blacks represented 20 per cent and Coloureds 80 per cent of the households with off-site sanitation.
Table 2 Sanitary facilities for urban and rural households in the Western Cape province, by race, 1995–99 (percentage distribution)
The proportion of urban Black households with off-site sanitation increased from 23,1 per cent in 1995 to 30,4 per cent in 1999, and of urban Coloured households from 1,1 per cent to 3,1 per cent. In rural areas, this proportion increased from 9,0 per cent in 1995 to 27,6 per cent in 1999 for Blacks and from 13,4 per cent to 19,6 per cent for Coloured households.
3.1.3 Clean water
The type of water available was classified into one of three categories: tap, borehole/tank, and natural source/other. The results in indicate that generally there is a decreasing trend for the percentage of households using dirty water from natural sources (i.e. well, spring and dam) or other sources (other than tap, borehole and tank). Overall, the percentage of households whose main source of water was natural sources or others decreased from 1,6 per cent in 1995 to 0,3 per cent in 1999 in urban areas and from 17,0 per cent to 4,9 per cent in rural areas.
Table 3 Main source of water for urban and rural households in the Western Cape province, by race, 1995–99 (percentage distribution)
3.1.4 Energy (electricity)
The number of households with access to electricity from the mains was calculated from answers to the question of which principal source of energy the household was using for lighting, since there are households which have access to mains electricity yet use other sources of energy for heating and cooking. shows that access to energy is not a major problem in the Western Cape province. The lowest percentage of households with access to electricity is consistently found among Blacks, but this seems to be declining over time. By 1999, two-thirds (64,6 per cent) of Black households had access to electricity from the mains.
Table 4 Availablity of electricity for lighting among urban and rural households in the Western Cape province, by race, 1995–99 (percentage distribution)
3.2 Participation rate, employment status, unemployment rate, education, economic sector and occupation of the head of household over time
3.2.1 Participation rate
The labour force participation rate was calculated by taking the number of household heads who were able and willing to work (including those who are actually working and the unemployed), as a percentage of the total number of people of working age.Footnote6 shows that participation rates for the different races have remained more or less constant over time, ranging between 70 per cent and 90 per cent for males and from 30 per cent to 70 per cent for females.
Table 5 Labour force participation among urban and rural heads of households in the Western Cape province, by race and sex, 1995–99 (percentage distribution)
3.2.2 Employment status
Employment status was classified into three categories: employed, unemployed, and non-economically active. ‘Non-economically active’ means either able but not willing to work, or disabled. As expected, rates of ‘non-economically active’ are higher among females than males, owing to family social and cultural obligations. This holds regardless of race and in both urban and rural areas.Footnote7 For urban areas, the proportion who were not economically active ranges between 14,2 per cent in 1997 and 30,9 per cent in 1998 for males, and between 12,9 per cent in 1997 and 56,2 per cent in 1998 for females. For rural areas, the corresponding figures are from 5,4 per cent in 1995 to 16,2 per cent in 1997 for males, and between 28,6 per cent in 1997 and 62,3 per cent in 1996 for females.
3.2.3 Unemployment rate
depicts how unemployment has been decreasing over time, especially for female heads of households. contains the percentages from which the figure was drawn. Overall, Blacks have the highest level of unemployment, while females of all races have higher unemployment rates than males. In urban areas, male unemployment rates ranged from 2,3 per cent for Whites in 1998 to 20,2 per cent for Indians in 1995, whereas among urban females unemployment rates ranged from 3,1 per cent for Whites in 1999 to 40,2 per cent for Blacks in 1995. Unemployment rates for males in rural areas were zero for both Whites and Blacks in 1997 but reached 19,3 per cent for Whites in 1996; unemployment was also zero for Coloured females in rural areas in 1996 and 1998 and was highest, at 25,3 per cent, for rural White females (in 1995).
Figure 2 Unemployment among Black and Coloured female heads of households in urban areas of the Western Cape province, 1995–99
Table 6 Unemployment among urban and rural heads of households in the Western Cape province, by race and sex, 1995–99 (per cent)
Over time there was an increase in the response ‘lack of skill’ cited by heads of households as the reason for their unemployment. shows the percentages who gave different responses to the question on why they were not working. A possible explanation is that after world-wide trade liberalisation, employers might demand relatively more skilled and less unskilled labour so as to produce better quality goods, capable of competing on the world market. Those household heads professing ‘lack of skill’ as the cause of their unemployment increased between 1995 and 1999, from 24,7 per cent to 37,5 per cent of the total for Black males and from 24,1 per cent to 27,4 per cent for Black females.
Table 7 Reason given for unemployment by urban and rural heads of households in the Western Cape province, by race and sex, 1995–99 (per cent)
3.2.4 Education
and show that Black and Coloured heads of households are concentrated at low levels of education, and that the situation is not improving. For example, the rate for heads with no education ranges between 8,4 per cent in 1997 and 10,8 per cent in 1996 for Black males, and between 6,7 per cent in 1996 and 13,2 per cent in 1999 for Black females. This is unlike the situation for Indians and Whites. For White males, the corresponding rates are zero (in 1995, 1998 and 1999) and 4,6 per cent (in 1996), and for White females also zero (1995 and 1999) and 4,8 per cent (1996). The Black female ‘no education’ rate increased from 8,8 per cent to 13,2 per cent over the time period. There is a decreasing trend among Coloureds for people with no education; the rate decreased from 6,2 per cent to 2,6 per cent for males, and from 12,5 per cent to 8,6 per cent for females between 1995 and 1999.
Table 8a Educational attainment of male urban heads of households in the Western Cape province, by race, 1995–99 (percentage distribution)
Table 8b Educational attainment of urban female heads of households in the Western Cape province, by race, 1995–99 (percentage distribution)
On average, 35,2 per cent, 22,5 per cent, 12,0 per cent and 3,6 per cent of the male heads of household have their highest education level as ‘primary school education or below’ for Blacks, Coloureds, Indians and Whites, respectively. For females, the rates are 31,7 per cent, 35,7 per cent and 4,4 per cent for Blacks, Coloureds and Whites, respectively.
3.2.5 Economic sector
There is little difference among the races in ‘type of economic sector’, except that more females than males work as domestic servants and that this does not seem to be changing over the years. The table of results can be obtained from the authors.Footnote8
3.2.6 Occupation
Occupation was classified into one of four categories: professional, clerical/service, skilled, and unskilled. While White and Indian heads of household predominantly occupy managerial or professional positions, Blacks and Coloureds are mainly in skilled and unskilled occupations.Footnote9 Coloureds are better off than Blacks. Both Coloured males and females are shifting away from ‘unskilled’ occupations and increasingly occupy managerial and professional jobs. Because occupation is a good correlate of education, lack of education is the probable reason why Blacks and Coloureds are clustered at the lower levels of the occupational ladder.
Changes of occupational category for Black heads of households seem to be far slower than those of Coloureds, if indeed there are any significant changes. The percentage of Black women (see ) in unskilled job positions increased from 41,5 per cent in 1995 to 76,3 per cent in 1999 in urban areas. This might however have reflected an increase in employment (see 3.2.3 above) which was concentrated at the lower occupational levels. For Coloured women, the percentage occupying unskilled jobs increased from 38,9 per cent to 47,8 per cent.
Figure 3 Occupations of black female heads of households in urban areas of the Western Cape province, 1995–99 (per cent)
3.3 Distribution of total household expenditure over time
and provide data on total monthly household expenditures. Comparable data could not be obtained from the OHS for 1995. It is noticeable from the tables that the numbers of Black and Coloured households spending less than R800 per month did not fall, even when the same nominal income intervals were used for the different years, considering the persistence of inflation. (If real incomes had remained constant, inflation would have pushed many households into higher nominal income brackets.) The OHS data demonstrate that the numbers with monthly expenditure below R800 are increasing, even in urban areas. It is inconceivable that changes in household size and demographics over the four-year period could play a part here. This is an indication that, in general, Black and Coloured households became poorer during the 1990s.
Table 9a Total monthly household expenditure by urban residents of the Western Cape province, by race, 1996–99 (percentage distribution)
Table 9b Total monthly household expenditure by rural residents of the Western Cape province, by race, 1995–99 (percentage distribution)
The percentage of urban Black households spending less than R800Footnote10 per month increased from 51,8 per cent in 1996 to 58,8 per cent in 1999, while for urban Coloureds the increase was from 17,3 per cent to 26,4 per cent. In rural areas the situation is somewhat better; and some alleviation of poverty may be occurring. For example, the percentage of Black households spending less than R800 per month in rural areas decreased from 97,7 per cent in 1996 to 67 per cent in 1999. However, one cannot be sure that poverty actually fell unless real per capita household expenditure shows the same trend.
For Whites, the percentage of households spending less than R800 per month in urban areas stayed more or less the same, ranging between 4,0 per cent in 1996 and 9,6 per cent in 1997. By 1999, it was 5,7 per cent. This, however, suggests that some Whites may also be becoming poorer.
The results indicate that Black and Coloured households spend far less than Indians and Whites. The typically larger size of Black households aggravates the situation.
4. Conclusion
The results of this empirical study have indicated that relatively large numbers of Black and Coloured households may have become poorer between 1995 and 1999, considering the substandard conditions in which they lived and their low levels of total household expenditure, which seemed to decrease over time. If this interpretation is correct, the Western Cape province will not come out of the chronic poverty that its people have suffered for decades unless something is done about educating Blacks and Coloureds. The statistics have also indicated that poverty levels vary with area of residence, race and gender. The most vulnerable are urban households headed by Black women, Black men, Coloured women and Coloured men, in that order. Seemingly, the hurdle of obtaining employment is not the most important issue. Rather, general levels of education of Blacks and Coloureds have to be increased so that they can climb the earnings ladder. Low levels of education are likely to be keeping them from obtaining an acceptable standard of living.
On the side of social capital, the study indicates that the government's challenge at this time is to do more for the poor. Relatively many Black and Coloured households, especially in urban areas, live without basic social services; for example, many urgently need decent housing and good sanitation.
It has been made clear that we are dealing with chronic poverty in the Western Cape province. Poverty eradication in the province might require a comprehensive long-term plan that will have to incorporate increasing the levels of education of Blacks and Coloureds in general. Short-term measures for better access to decent housing and good sanitation might also be necessary to deal with the deplorable living conditions of the poor, especially those in informal setlements in urban areas. These two strategies are likely to help in ensuring acceptable standards of living for all households in the province.
4.1 Limitations of the study
Earlier we discussed issues relating to the difficulty in combining information from the OHS for different years. Changes in the wording of questions, number of response options, and specific response options, make it difficult to assure that a variable is measuring the same quantity from year to year. Further, respondents may not answer questions honestly, or may provide the survey enumerators with inaccurate guesses rather than precise answers.
4.2 Further research
To investigate the dynamics of poverty in the Western Cape province in a rigorous manner, further research is necessary. The specific objectives that should be pursued are: to find the level of absolute poverty in the Western Cape province; to find the level of income inequality in the province; to identify the factors underlying poverty in the province; to distinguish the effects of migration and AIDS on poverty in the province; to identify the particular constraints of poverty alleviation on a local government level; to formulate the socio-economic strategies that the government can use to eradicate absolute and relative poverty in the province; and finally to make recommendations on how best the government can deal with the poverty problem in the Western Cape province.
Additional information
Notes on contributors
R Madsen
University of the Western Cape and University of Missouri. We thank the University of the Western Cape for financial assistance.Notes
2State of the Nation address of the President of the Republic of South Africa at the opening of Parliament, Cape Town, 9 February 2001.
3 Statistics South Africa is the government agency which publishes the country's official statistics.
4In this paper, ‘Black’ is used for ‘African/Black’ and ‘Indian’ is used for ‘Indian/Asian’.
5The percentage figures discussed in this section come from the tables. They were calculated from the OHS data.
6In South Africa people start working legally from the age of 15.
7For detailed tables on employment status over time (urban and rural), the authors can be contacted for a copy of the Technical Report, UWC-TRB/2003-02. These tables are not included for the sake of brevity.
8Some tables and figures are not included in the paper because of space limitations.
9For a detailed table on occupation over time, the authors can be contacted for a copy of the Technical Report, UWC-TRB/2003-02. This table is not included for the sake of brevity.
10R800 per month was the official poverty line for households in South Africa in 1999.
Related Research Data
References
- Bhorat , H. , Leibbrandt , M. and Woolard , I. 1995 . Paper presented at the African Economic Research Consortium's Conference on Transitional and Long-Term Development Issues . 1995 , Johannesburg. Towards an understanding of South Africa's inequality. , 30 November .
- Coetzee , Z. R. , Gwarada , K. , Naudé , W. and Swanepoel , J. 1997 . Effect of tariff reduction on employment and output in industries , Potchefstroom University .
- Deaton , A. 1998 . The Analysis of Household Surveys: a microeconomic approach to development policy , Baltimore and London : Johns Hopkins University Press .
- Freeman , R. 1995 . Are your wages set in Beijing? . Journal of Economic Perspectives , 9 : 15 – 32 .
- Greene , W. H. 1993 . Econometric analysis , New York : Macmillan Publishing Company .
- Johnson , G. E. 1997 . Changes in earnings inequality: the role of demand shifts . Journal of Economic Perspectives , 11 ( 2 ) : 41 – 54 .
- Knight , J. B. and Mcgrath , M. D. 1987 . The erosion of apartheid in the South African labour market: measures and mechanisms , Oxford : Oxford University Institute of Economics and Statistics . Applied Economics Discussion Paper no. 35, Institute of Economics and Statistics
- Moll , P. G. 1995 . Discrimination is declining in South Africa, but inequality is not , Unpublished article
- National Research Foundation (NRF) . 2000 . A guide to NRF research support for 2001 , Pretoria : National Research Foundation .
- Naudé , W. and Serumaga-Zake , P. 2001 . An econometric analysis of labour force participation and unemployment in the South Africa's North-West province . Development Southern Africa , 18 ( 3 ) : 261 – 78 .
- Richardson , J. D. 1995 . Income inequality and trade: how to think, what to conclude . Journal of Economic Perspectives , 9 : 33 – 55 .
- Serumaga-Zake , P. and Naudé , W. 2001 . Labour market discrimination in the North-West province . Development Southern Africa , 18 ( 5 ) : 671 – 82 .
- Serumaga-Zake , P. and Naudé , W. 2002 . The determinants of rural and urban household poverty in the North-West province of South Africa . Development Southern Africa , 19 ( 4 ) : 561 – 72 .
- Wharton Econometric Forecasting Associates (WEFA) . 1996 . South African competitiveness monitor , Volume 4 , Pretoria : WEFA Group . Provincial Comparisons
- Whiteford , A. , Posel , D. and Kelatwang , T. 1995 . A profile of poverty , Pretoria : HSRC .
- Wood , A. 1995 . How trade hurt unskilled workers . Journal of Economic Perspectives , 9 : 57 – 80 .