C&S=CD-Rom97=City, Citizenship and Gender
Chapter: 02 Equal Opportunities
Document: EO-2.3.03
Kb: 201 Words: 27.542
En (English-Anglais)

+++++++++++++++++++++++++

September 1983

--------------------

Dori Posel
Department of Economics, University of Natal, Durban
Alison Tales
Department of Town and Regional Planning, University of Natal., Durban
Michelle Friedman
National Land Committee

-----------------

Introduction
1. Women and gender in region E's "formal" economy
1.1 Rates of participation:
1.2 Women's sectoral employment and occupational segregation
1.2.1 Employment by sector:
1.2.2 Employment by occupation:
1.3 Employment opportunities, wages and access to income

Women in the informal sector
2.1 The informal sector and income-earning opportunities:
2.2 Deregulation and gender relations:
2.3 Domestic work, childcare responsibilities and informal work:
2.4 Urban - rural differences in income-earning opportunities:

Women in the rural economy of region
3.1 Subsistence agriculture
3.2 Smallholder farming:
3.3 Commercial farming:
3.4 Income generation oriented development projects:
3 5 Concluding comments:

Social reproduction and access to resources
4.1 Housing:
4.2 Water:
4.3 Energy:
4.4 Transport:
4 5 Health:
4.6 Educare:
4.7 Concluding comments:

INTRODUCTION

Introductory Comments:

When pointing to "glaring levels of inequality" (Harrison, 1993) in South Africa as a whole, and in region E specifically, most studies refer to racial inequalities in access to political representation, basic resources, employment opportunities, and rates of remuneration. Inequalities between men and women, within each race group, and across races, are rarely identified. And yet, as this study will illustrate, there are significant differences between men's and women's socioeconomic position in region E.

The complexity of gender differentials often is only revealed by asking the "right" questions of economic and social processes. These questions originate through recognising "gender", as a category of analysis, similar to, and as important as, that of race (or class). Gender refers to the ways in which biological differences between men and women have been translated into social differences. At the level of economics, gender appears as a sexual division of labour, where some types of work are associated with women, and others, with men. Broadly, production is seen to be men's work, and reproduction, the responsibility of women.

If the different rates, and nature, of economic participation between men and women are "naturalised", or taken as biologically inevitable, then the complex ways in which the sexual division of labour are reinforced and perpetuated, will be invisible. Women's disproportionate involvement in unpaid and paid domestic work, and community services, can be attributed simply to their sex-specific characteristics. As insidious as this suggestion might be to explain and justify racial differences in access to resources, most studies that ignore questions of gender implicitly are subscribing to, (and possibly unintentionally) perpetuating, this view of sex differences.

The argument that there are complementary roles for men and women ignores the reality that the costs and benefits of the sexual division of labour are not shared equally among men and women. The range of choices available to women, and the rewards to their work, are not equal to those of men's. Although more women are entering paid employment, women are still viewed as secondary wage earners with primary responsibility to the home, a division of labour which has significant implications for women's earning potential in the formal labour market, and for the kinds of jobs and benefits that are available to them.

This is not to suggest that all women are equally disadvantaged: race and gender intersect in important ways in South Africa. White women, for example, have had greater access to resources and opportunities, as whites, than have African women. Many studies have documented the extent, and implications, of "race differences" in the economy, but very few have sought to explain this interplay between gender and race. Some theorists, such as Pillay (1985), have suggested that "racial discrimination should be regarded as the mayor stratifying element in the labour market", implying that gender discrimination would constitute a secondary form of stratification. Even if this suggestion were to be accepted, however, it still does not establish the nature of the intersection between gender and race, and the extent of their interdependence.

Internationally, attempts to explore the interplay between gender and race have been debated extensively. It is not the intention of this report to contribute directly to this analytical debate, nor to redress the theoretical lacunas in South African literature. Rather, the study seeks to identify and document gender differences between and within race groups in one region in South Africa, and to point to implications of these differences for development in the region. The examination of why there are (sometimes huge) inequities between men and women, and amongst women, should not be read as comprehensive or exhaustive. Rather, such analysis should be the subject of further extensive and extended research.

Outline:

The study is divided into five sections. The first three sections examine women's participation in paid work, and consider the main characteristics of women's work in the formal, informal and rural economy in region E.

Much of the work which women do, however, is not paid. Women's involvement in unpaid work (in the home and in subsistence farming) places considerable constraints on their access to income-earning opportunities. Yet, unpaid work is largely ignored by economists and policy-makers, and is treated as a "natural" task or commitment which women will, and should, undertake. The nature of women's unpaid work is influenced by the kinds of resources which are available to women. Section 4 of the study examines these issues) and considers unpaid work in the context of women's access to water, electricity, housing, transport and childcare facilities. It also considers the interplay between women's access to resources and their position in the economy more generally.

Section 5 summarises the main features of women's socioeconomic position in region E, and attempts also to point to future trends in the region. In particular, the nature of certain policies, which are being debated for the region, are re-examined from a gender perspective and their possible impact on women is considered.

Data Sources:

In exploring the position of women in region E, the study has drawn on those official statistics that are available, making extensive use of census data. Although there are many problems with these data. they remain a relatively useful source for identifying aspects of women's socio-economic position in the region, and particularly, for a cross-sectional analysis of gender and race differences. Census data are also one of the only sets of statistics which provide regional and national figures.

The limitations of census data, however, need to be considered carefully, and recognised as important constraints on an empirical investigation. The comparison of census results is practically difficult and limited, thereby complicating attempts to examine changing trends and patterns. Both the 1980 and the 1991 census results were adjusted for under-enumeration. The figures in the 1985 census, however, were not. Changes in the nature and extent of women's employment and access to income over the last ten years therefore must be inferred only from the 1980 and 1991 figures.

Even comparisons of these two data sets, however, are problematic. The nature of the various sub-divisions that have been used in the census has changed between 1980 and 1991. Over the last ten years, for example, the boundaries of districts have been altered, complicating an analysis of inter-regional trends. The classification of occupations has also changed, and a new occupational category, "Transport, delivery and communications" has been introduced.

These kinds of changes make it difficult to compare actual numbers across the census and to quantify changes in absolute terms. It is only possible to use time series analysis to infer relative, percentage changes in women's participation in the economy since 1980.

Census data are limited not only because of the difficulty of comparing figures, but also because the figures do not reflect all of women's socio-economic activity. Although women constitute more than half the population of region E, the 1991 census suggests that only approximately two fifths of women had paid employment. The remainder of women either work in the informal sector or in the home. Neither of these work activities is adequately quantified, if quantified at all.

Micro-studies illustrate that many women are involved in work in the informal sector. Employment and earnings in the informal sector, however, are not specified, but are included in the "unspecified and not economically classified (NEC) category". This category also includes: those whose occupations are not specified or are not included in the census breakdown of occupations; and those who are unemployed and whose occupations are unspecified. In South Africa as a whole, half of all people listed as "occupation unspecified and NEC" are women (52.7%) and of these, 83.5% are African women.

Many aspects of women's economic activity are not quantified at all. Data for the census is collected on the assumption that only work which is paid is economically productive, and that only paid workers are economically active. Subsistence farming and private domestic labour (work undertaken largely by women) therefore are not viewed as productive work, and women involved in these activities are classified as "not economically active".

A thorough examination of women's socio-economic position in region E is clearly hampered by the absence of consistent and comprehensive data. There are also dimensions of gender relations which are not amenable to quantification. These factors necessitate that quantitative data be complemented by qualitative case studies, where possible, which examine gendered activities and gender dynamics not recorded or not possible to quantify.

Finally, it should be noted that the Transkei part of region E has been excluded from this analysis. The detail that is necessary for the study is not available in the Transkei census results, and the South African census excludes the TBVC states. The study focuses, therefore, on Natal and KwaZulu.

1. Women and gender in region e's "formal" economy

The position of women in region E is similar to the position of women in South Africa as a whole, although women form a larger proportion of the total population in the region than in South Africa. This section examines women's participation in the formal (paid) economy in region E, and considers differences between men and women, and amongst women, in rates of employment, kinds of employment and access to income.

Much of this analysis is drawn from census data. As noted above, these figures have limitations, not least of which is that they can desensitise, and even anaesthetise, the reader. One of the more specific problems that arises when trying to quantify women's employment is the census category of "economically active", which includes those currently employed at the time of the census, as well as those frictionally or "temporarily" unemployed. This broad classification creates difficulty when trying to isolate levels of employment and unemployment. Some percentage of the unemployed are probably reflected in the "unspecified" sub-divisions, but this figure will only represent those unemployed workers whose (previous) employment by sector or occupation has not been specified, or which is not aligned with the census breakdown. For reasons of brevity and simplicity, however, figures on economically active will often be treated as "employment" statistics.

These qualifications aside, the census data remain a guide, and possibly the most "complete" enumerative guide from which positions and patterns in women's employment can be deciphered and decoded.

Rates of participation:

In 1991, women formed 52.5% of the total population in region E, but accounted for only 42.9% of all those classified as economically active.

Table 1: Women as a percentage of the total population

----------------------------------------------------------------

SA as a whole Region E
---------
1980 49.1% 51.7%
1991 50% 52.5%
----------
Table 2: Women as a percentage of the economically active workforce
-----------
SA as a whole Region E
-----------
1980 32.3% 33.7%
1991 39.4% 42.9%
---------

source: CSS reports: 03-80-08 (1980); 03-01-08 (1991)

Although a smaller proportion of women than men in all race groups were engaged in waged work, participation rates for women also varied across race categories. The highest rates were amongst African women, and the lowest, amongst Indian women.

Table 3: Participation by gender within race groups

Percentage who were:

1980 White Coloured Indian African

Women 34.4% 41.1% 25.5% 34.8%

Men 65.6% 58.9% 74.5% 65.2%

1991

Women 39.6% 44.0% 32.0% 45.2%

Men 60.4% 56.0% 68.0% 54.8%

As illustrated in Table 3, 45.2% of all Africans classified as economically active were women. Tables 4 and 5 illustrate that African women also formed the largest proportion of all women (82.2%) and of all economically active women (78.4%). Participation in the paid economy was most skewed for Indian women, who accounted for only 32% of all Indians classified as economically active.

Differences in women's participation by race requires explanation, although it is not possible to do that here. Economic pressures, social support networks, gender relations, as well as "cultural" and ideological factors may all be important.

Table 4: Population in region e by gender and race

Percentage who were:

1980 White Coloured Indian African

Women 9.3% 1.5% 10.6% 78.6%

Men 9.4% 1.5% 11.2% 77.9%

1991

Women 7.2% 1.3% 9.3% 82.2%

Men 7.9% 1.4% 10.0% 90.7%

source: CSS reports: 02-80-05 (1980); 03-01-01 (1991)

Table 5: Economically active in region e by gender and race

Percentage who were:

1980 White Coloured Indian African

Women 14.9% 2.3% 9.5% 73.3%

Men 14.5% 1.6% 14.1% 60.8%

1991

Women 10.7% 1.9% 9.0% 78.4%

Men 12.3% 1.8% 14.5% 71.4%

source: CSS reports: 02-80-05 (1980); 03-01-01 (1991)

Trends since 1980 in Participation by Gender sad Race:

One of the most notable changes in economic participation by gender and race has been the increase in women's involvement in region E's economy across all race groups. This increase has exceeded women's increasing employment in South Africa as a whole. In 1980, 32.3% of all people classified as economically active in South Africa were women. In 1991, this figure rose to 39.4%. Between 1980 and 1991 in region E, an additional 9.2% of all economically active people were women, representing a relative increase in women's "employment" of 27.3%.

This overall increase, however, was distributed unevenly across race groups:

• The most significant change was amongst African women: according to the 1991 census, 29.9% more African women were recorded as being economically active in region E than in the 1980 census. African women as a proportion of the total population increased from 52% in 1980 to 53% in 1991. This alone, however, does not account for the change in African women's employment. Rather, as will be discussed later, the trend reflects the increased migration of African women into urban areas, as well as the increased employment opportunities for African women in these areas.

• Although considerably more Indian men than women are economically active, the gap has narrowed over the last ten years. Indian women's participation in paid work increased from 25.5% of all economically active Indians in 1980, to 32% in 1991, representing a substantial increase of 25.5% in the number of economically active Indian women.

• 16% more white women were classified as being economically active in 1991 than in 1980. White women as a proportion of all economically active whites increased from 34.4% in 1980 to 39.6% in 1991.

• The smallest increase in economically active women was amongst coloured women, whose involvement in paid work increased by only 9.5% over the past ten years. A relatively large percentage of all coloureds classified as economically active, however, were women (41.1% in 1980 and 44% in 1991).

Analysing these Trends:

The increase in women's employment in region E, and in South Africa as a whole, follows international trends of an increasing feminisation of the labour force. During the decades that women's participation in the formal economy has increased, the rate of men's participation (world-wide) has slowly decreased, such that the gap between the two has narrowed.

It is not possible to say with certainty why the labour force has become more "feminised". Indisputably, recessionary conditions have placed more pressure on women, in particular, to find employment. Within this context, some explanations have focused on supply-side factors, and have pointed to women becoming more work oriented, increasing their levels of education and training, and challenging their "domestic roles". These factors, however, may be as much an effect of women's increasing participation in the paid economy, as a cause. Furthermore, although education and work attitudes affect access to employment, supply-side explanations assume also that labour can choose when to be employed and in what capacity, a relatively unrealistic assumption for most workers.

Perhaps a more convincing explanation of the increased employment of women is one which rests on demand-side factors. Whilst this explanation does not dismiss supply-side considerations, it is argued further that employment depends ultimately on jobs and markets being, and becoming, available. The feminisation of the labour force is then explained by examining technological and organisational changes in the economy, and by interpreting how these changes have affected the labour process and the employment of women. Three changes which have been highlighted are the growth in the public sector, changes in the structure of production, and the increasing importance of the tertiary sector in the economy.

The public sector tends to be labour-intensive and has been a mayor source of job opportunities for women. A 1982 report released by the Organisation for Economic and Co-operation Development (OECD) identified that the public sector in OECD countries had become more female intensive over time. The report also identified a positive correlation between growth in the public sector and the feminisation of public sector employment. Although the reasons for women's employment in the public sector will be explored more extensively in the next sub-section, Hagen and Jenson offer an immediately accessible explanation when they observe that women are employed in public sector jobs "because so many of them were in the area of care-giving and social services long associated in popular discourse with 'feminine talents'" (1988, 7).

The growth of the public sector was associated with the welfare state. In the 1970s and 1980s, however, many countries cut back their public sector. Employment in social services was particularly affected, although these jobs did not always disappear as some public sector work was privatised. As will be noted below, despite talk of privatisation in South Africa, the public sector, and women's employment in public services, has grown in the region.

The increase in women's employment internationally has also been attributed to changes in the nature and structure of production in many economies. Attempts to increase the competitiveness of industries led to cost-cutting and moves towards increasing flexibility in the production process, thereby enabling production to respond to changing market conditions. In the 1960s and 70s, some industries (or parts of them) moved to peripheral areas and to the "third world" inter alla in search of cheaper labour. In many cases, women were employed.

In the 1970s and 80s, more fundamental changes in the production process have occurred in developing, and especially industrialised, countries. These changes have included the introduction of new technologies and forms of organisations, and have necessitated a more flexible labour force, precipitating further increases in the employment of women in some industries. As Hagen and Jenson have argued:

Women have emerged as very desirable employees in these circumstances because their relationship to the labour market traditionally displayed the characteristics of flexibility ... Individual women have had a less continuous relationship to the labour market, moving in and out of employment in response to the demand for labour and their personal possibilities (1988, 10).

The 1980s has also seen the growth of small business internationally, and women's employment in these highly competitive undertakings. The perception of women as cheaper, more flexible. and more acquiescent workers, goes some way towards accounting for the employment of women in these undertakings.

The relative decline of manufacturing, and the growth of the tertiary sector (commerce, finance, personal services etc) are further important processes in the feminisation of the labour force. Most economies, but especially the industrialised, countries, have experienced growth in the tertiary sector. This sector has displaced manufacturing as the mayor form of employment in many industrialised, economies. Many occupations within the tertiary sector are seen as drawing on women's "natural" abilities and attributes and have come to be seen as women's work. Occupations such as clerical work, cleaning, reception, sales, and service, for example, are generally occupied by women. In the 1980s, parts of the tertiary sector have also experienced the shift towards flexibilisation as discussed above. For example, more work is being done from the home and the proportion of part-time and casual workers has risen, trends which have precipitated the increased employment of women.

In many respects, the increasing feminisation of the labour force in region E can be explained by processes which have paralleled international experiences. It will be shown that the two biggest employment sectors of women are "community, social and personal services", which is largely public sector employment, and manufacturing. Furthermore, as the community service and manufacturing sectors have expanded, so women's participation in these sectors has increased. Two other important sectors of women's employment are the commerce and finance sectors, which although smaller, together account for a similar number of women to those employed in manufacturing.

One notable difference between employment trends in South Africa and those internationally is that moves towards greater flexibility and competitiveness in domestic production have been constrained by South Africa's limited access to new technology. It is quite likely, however, that any future restructuring of production, which necessitates a flexible labour force, will lead to an increase in women's employment in these industries.

1.2 Women's sectoral employment and occupational segregation

One of the most important insights of demand-side explanations of the feminisation of the labour force lies in the suggestion that although more women are entering paid work, certain traditional assumptions about the kinds of work women should do have not been challenged fundamentally. The position of women in region E cannot be inferred, therefore, simply by the extent of women's participation in the economy (Budlender, 1991). Such correlation would ignore the nature of women's employment and their access to income-generating opportunities.

Furthermore, any consideration of future trends and development in the region must consider how policies might impact on women s employment. The effects of privatisation cuts in public expenditure and the restructuring of production on women's employment, for example, can only be inferred from identifying where women are presently employed, and why their employment has been increasing.

Although women's participation in paid work has increased between 1980 and 1991, it has not risen proportionately in all sectors and in all kinds of employment. Rather, women have moved into specific sectors, and within these sectors, into specific kinds of jobs. The labour market in region E, and in South Africa as a whole, reflects a crowding of women into certain sectors and an occupational segregation between men and women within sectors.

1.2.1 Employment by sector:

According to the 1991 census (and as reflected in tables 6 and 7), the biggest employment sector of women, is "community, social and personal services". More than a third (34.52%) of all economically active women were classified in this sector. "Community, social and personal services" is also the largest employment sector (accounting for 510 844 of all economically active) and the only sector where women predominate over men: women employed in this sector outnumbered men by almost 2:1.

The second largest employment sector of women is manufacturing, which accounted for 10.9% of all women's employment in 1991. Manufacturing remains a sector which is dominated by men, however, where more than two thirds (67.5%) were men, and only 32.5% were women. Women's participation in the "agriculture and related" sector is similar to that in manufacturing. Agriculture is a much smaller employment sector, however, and accounted for only 7.1% of women's employment.

Women's rates of participation in the "wholesale, retail, catering and accommodation" (commerce) sector, and in the "financing, insurance, real estate and business services" (finance) sector largely match women's participation in region E as a whole. Of all those classified as economically active in commerce, 38.8% were women, and in financing and related, 41.7% were women. These two sectors together accounted for the employment of 13.9% of all economically active women, higher than manufacturing.

Highly skewed employment sectors are mining and quarrying, where women represented 5.5% of all workers in this sector; construction, where 6.4% of those employed were women; and "electricity, gas and water", where 8.9% were women. Of all economically active women, 0.2% were "employed" in mining and quarrying; 0.64% in construction; and 0.13% in electricity, gas and water.

The largest percentage (35.1%) of all economically active women in 1991 were classified in the category "activities not adequately defined".

Although it is difficult to isolate the various components of this category, this 35.1% includes women who were unemployed at the time of the census and whose (previous) sectoral employment was not defined adequately, and women engaged in informal work.

1. The 1991 breakdown for sectors did not exactly match the 1980 breakdown. Additions to the 1991 census classification are noted in brackets in the table.

Differences Amongst Women:

An analysis of women's employment by sector not only identifies differences between men and women, but also differences amongst women:

• Of all economically active white women, 31.5% were concentrated in the service sector, 19.5% were classified in the financial sector, and 17.7% were found in wholesale and related. Only 9.6% of all white women workers were listed in sectoral employment "not adequately defined".

• Most coloured women were involved in the wholesale and related sector, which accounted for 27.3% of all economically active coloured women, and in the service sector (25.6%). Nearly a quarter (24.5%) of all coloured women workers were classified in the "not adequately defined" category.

• Nearly a third (32.3%} of all economically active Indian women were employed in the manufacturing sector. Approximately one fifth (19%) of all Indian women workers were involved in the wholesale and related sector, and 18.8% were in the service sector. One fifth of all economically active Indian women were listed in the unspecified category.

• African women were concentrated in the service sector, which accounted for just under a third (32.5%) of all economically active African women, in the agricultural and related sector (8.8%) and in the manufacturing sector (8.4%). Although African women formed 71.4% of all economically active women, only 0.7% of African women workers were involved in the financial sector, and 8.3% in the wholesale and related sector. Two fifths (40.5%) of all economically active African women were involved in activities which were "not adequately defined".

These differences amongst women begin to shed some light on the relationship between race and gender, and the extent to which these have influenced women's access to certain kinds of employment. A relatively high percentage of all women were concentrated in the service sector, but women's involvement in the financial, wholesale, manufacturing and agricultural sectors varied by race.

White women were over-represented in the financial sector, a high-income sector, but less than one percent of all African women workers were found in this sector. White, Indian and coloured women were also relatively well represented in the wholesale and related sector, whereas less than a tenth of all economically active African women were employed in wholesale and related activities. Although census data does not include subsistence farming in the agricultural sector, the figures still indicate that African women formed the largest percentage of all women involved in agriculture. Furthermore, of all women whose employment was not adequately defined, the majority were African women, indicating the extent to which African women are involved in informal sector activities.

Differences amongst women will become more apparent when the nature of women's employment within each sector is examined.

Trends and Shifts in Sectoral Employment:

When accounting for the increase in women's employment in region E between 1980 and 1991, a primary question to be considered is where women have been incorporated into the economy. Have women entered sectors that are growing, or has the gender composition within contracting, or largely stagnating sectors changed? In other words, are women being drawn into the labour market as new jobs are created, or are women replacing men in existing jobs?

• According to the census results, the number of economically active people classified in community, social and personal services has increased by 43.7% over the last ten years. Of all new entrants into this sector, an overwhelming majority (86.4%) were women, 3 and of these women, 88.1% were African women. This sector, therefore, has become more female intensive over the last ten years: in 1980, 52.9% of all service sector workers were women; in 1991, this proportion increased to 63.1% (see activities).

• From 1980 to 1991, the number of economically active people classified in manufacturing, the second largest employment sector of women, has increased by only 10%. 98.7% of all new entrants, however, were women, and of these) 85.7% were African women.

• Since 1980, employment in "financing, insurance, real estate and business services" has increased by 88%. Less than half (43.7%) of all new entrants in this sector were women, and of these, 48.6% were white, and 25.9% were Indian women.

• Employment in the "wholesale, retail, catering and accommodation" sector has increased by 45.9% since 1980. Of all new entrants into this sector, approximately 46.6% were women, the majority of whom (63.5%) were African women, and Indian women (24.7%).

• There has been a significant increase in the number of economically active workers whose sectoral employment is listed as not adequately defined. In 1991, this category accounted for 717 500 people, the biggest "employment category" in the sectoral breakdown. Since 1980, the number of all workers classified in this category has increased by an enormous 218%. More than half (53.2%) of this increase represented women whose activities were not adequately defined, and of these women, 89.7% were African women.

• Since 1980, employment in the "agriculture and related" sector has declined by 11.4%. Women's employment in the sector, however, has increased from 27.7% to 32.6% of the total, while men's employment has declined in absolute terms.

• Employment in some of the highly male dominant sectors has declined or has been static between 1980 and 1991. Harrison (1993) shows that mining has declined at a rate of 3.3% pa, while transport and construction have grown at an annual rate of only 1.8% and 1.9% respectively.

Analysing these Trends:

Census figures highlight three of the more important trends which account for the increase in women's employment in region E: the expansion of the tertiary sector, and in particular, of the community service sector within it; women's increasing involvement in the informal economy; and the increased employment of women in manufacturing.

1. The growth of the tertiary sector has been associated with an increase in women's employment. Two different processes account for this trend:

Firstly, commerce and finance, which employ a relatively large proportion of women have experienced strong growth in the 1980s. Women's participation in these sectors has grown, although the increase is slight. This trend is similar to the international experience. Unfortunately, we have no evidence on how the structure of the labour process in these sectors is changing and the implications of these changes for the nature of women's employment.

Women in the finance sector are largely white, and to a lesser extent Indian. The growth of this sector has not contributed to the growth of African women's employment. Rather, African women are a larger and increasingly important part of the commerce sector (although they are far better represented in some other sectors). Their presence in the commerce sector probably reflects the large numbers of relatively low paid unskilled jobs in the retail and in tourist related activities (ea. as cleaners, in catering etc.). Many of the latter are likely to be temporary or part-time jobs.

Secondly, census figures indicate that one of the most significant factors accounting for the growth of women's employment - and especially of African women 's employment - has been the growth of the community services sector. A large part of this significant increase in women's employment is likely to be the result of the growth of the public sector.

Contrary to expectations that state policies favouring privatisation would lead to cuts in public sector employment, the development of the bicameral parliament, and of homeland governments has meant the reverse. Although there are still huge discrepancies in state expenditure on government services by race, pressure on the state has meant that the gap has narrowed somewhat in the 1980s. Further, the growing population has implied that facilities like schools have had to be expanded, although they have not necessarily kept pace with population growth and are by no means adequate.

The community services sector also includes domestic work. Unfortunately, available statistics do not allow us to see whether domestic work has grown or declined in the region in the 1980s. Given that domestic work is a significant employer of women, this is an important area for future research.

2. In the absence of a separate census classication for the informal sector, it is not possible to quantify exactly women's involvement in informal work. Many case-studies of informal work (which will be referred to in the subsequent section), however, indicate a significant increase in women engaged in informal activities. This increase may be attributed largely to the depressed state of the economy which has pushed more and ore women into self-employment and other survival strategies. It is therefore quite likely that the dramatic increase in employment which is listed as not adequately defined reflects an equivalent increase in women's work in the informal sector.

3. Although the increase in the manufacturing sector has been relatively low, most of the new employment in this sector represents the employment of women. When accounting for the growth of women's employment in manufacturing in the 1980s, two trends are of importance:

Firstly, there has been a growth of some 28% in women's employment in manufacturing in Durban, while men's employment grew by 2% in the same period. The participation of white and coloured men in manufacturing has declined, while there has been a growth in the employment of African women, and to a lesser extent of African men and Indian women. Although the reasons for these shifts require more investigation, it is quite possible that they have paralleled attempts to reduce costs in the context of recession in the 1980s. Studies have also suggested that there has been a rapid growth of small industry in Durban in the 1980s (McCarthy, 1993) which has been associated with the increased employment of women.

The growth of women's employment in Durban is probably also associated with subsectoral shifts within manufacturing. Unfortunately, the gender breakdown within manufacturing sectors is only available at a national level, making it impossible to determine regionally specific intrasectoral shifts. At a national level, however it is clear that some industries, primarily clothing, and to a lesser extent textiles and food, are female employing, while most others are overwhelmingly male employing. Between 1979 and 1985, important male dominant sectors, such as chemicals, metal and machinery, reduced employment (often in association with processes of capital intensification) while more female-employing industries, like clothing and food, grew. The growth of small industries also has often been in these female dominated subsectors, although not exclusively. (McCarthy, 1993).

Secondly, Bonnin et al (1991) suggest that industrial decentralisation policy has played a critical role in generating jobs for women. Between 1982 and 1991, some 53925 jobs were created under the (Regional Industrial Development) RIDP programme (Baldogh, 1993). In certain decentralisation points, as Table 8 shows, the growth of women s employment was extremely rapid, although from a low base. Women s employment in decentralisation points is largely confined to the KFC Industrial Estates. Decentralisation points in Natal, in contrast, employ a high proportion of men, presumably as a result of the earlier emphasis on heavy industry in industrial decentralisation policy. The growth of these areas has been much slower than in the KFC areas in the 1980s. In all cases, the growth of women s employment was considerably higher than the growth of manufacturing employment as a whole

Employment within the KFC estates grew from 6954 in 1980 to 41140 in 1990 (KFC, nd.). Some 57.5% of jobs in these points were occupied by women in 1990. The proportion of women employed is much higher in the more peripheral Ezakheni (67%) and Madadeni (86%) industrial points than in the more established and successful Isithebe (45%).

As is the case in Durban, the growth of women's employment in decentralised industries is partly tied to the importance of "traditional" women's industries. The clothing industry accounts for 32% of jobs in the (KFC, Estates, and for as much as 37.5% in Madadeni. The (KFC, points, however, appear also to have been associated with some gender shifts in terms of the dominant South African pattern. Plastics and footwear, both important industries in the estates, employ mainly women. Nevertheless, the KFC assures us that "females are mostly employed by light industries ... with heavier industries ... employing mainly males" (KFC, nd: 5).

In her study of decentralised manufacturing industries, Ardington (1984) accounts for the increased employment of women by suggesting also that women are prepared to work for lower wages and that, with fewer alternatives, they tend to be more willing to accept temporary employment. The employment of relatively inexperienced African women in these industries therefore represents an opportunity for industries to keep costs low in an increasingly competitive recessionary context.

Much of women s employment in decentralisation points typifies low wage employment. As Pudifin and Ward's study (1986), illustrated, for example, employment in these industries has been characterised by primitive working conditions, reports of sexual harassment, and a 30% wage differential between decentralised and metropolitan jobs. Nevertheless, employment in these industries is important for women. It represents a better alternative than farm work or domestic employment in some areas, and allows women to maintain closer ties with their families (Bonnie et al! 1991).

Table 8 : Growth in african women's employment decentralisation points and Durban, 1980, 1991

Women’s Employment Increase 1991: Women

Area in Manufacturing % as a %

1980 1991 of all in manufact.

Pietermaritzburg 3803 6 800 79% 28.2%

Ladysmith/

Ezakheni 5 562 9 267 67% 58.5%

Newcastle/

Madadeni 3 553 8 330 134% 36.6%

Richards Bay 1 748 2 262 29% 20.0%

Yryheid 345 767 122% 27.5%

Isithebe 1 159 3 292 184% 36.6%

Durban 56 377 72 141 28% 31.7%

Source: CSS Reports : 1980, 1991, KFC (nd.), Baldogh (1993)

Notes:

1. In most cases, other than Pietermaritzburg and Durban, the non African component of women’s manufacturing is very small. Notably in Pietermaritzburg and Durban, the growth of Indian women in manufacturing was also rapid.

2. Figures for Isithebe in the census data are low compared to other sources, so (KFC, data has been used.

3. The numbers employed in Madadeni are small (2691), accounting for the difference in the proportion of women employed.

Different periods of industrial decentralisation policy have been associated with varying implications for men's and women's employment.

The policies of the 1980s facilitated the growth for women s employment.

It would be interesting to see what implications the shift since 1991 has had in gender terms. For, this shift is clearly favouring "better located" areas - generally areas with a lower proportion of women employed.

Changes in Men's Employment:

Census figures do not establish the extent to which women are replacing men in existing jobs. They do indicate, however, that as new jobs are being created (particularly in the manufacturing and service sectors), so the employment of women, relative to that of men's, has increased. Part of the growth in women's employment can be attributed to the fact that women are "cheaper" to employ than men. It is not accidental, however, that the increased employment of women has paralleled the growth in service, clothing and textiles industries. These industries traditionally have employed women not only because women are prepared to work for lower wages, but also because the kind of jobs which are available are seen to be "women's work".

The feminisation of the labour force, however, cannot be seen only in the context of the growth of women's employment. The stagnation or decline of some of the traditionally male-employing industries is also an important factor. As we have noted, men's employment within manufacturing has stagnated, while mining has experienced significant declines. Construction and transport have grown very slowly in the period. Further, census figures indicate a gender shift in agriculture, along with an overall decline in employment in this sector. The gender shift here is likely to be associated with a shift to part-time and seasonal labour as well as with general cost cutting, which might explain the relative increase in women's employment in agriculture.

The growth of women s employment largely in urban areas, and the stagnation or decline in men's employment in the region, coupled with declines in important male dominant sectors, such as mining, outside of the region, probably underpins urbanisation trends noted elswhere (see May, 1993b; Todes and Smit, 1993). In the 1980s, women have begun to migrate from rural to urban areas, while men seem to be returning to rural areas.

Finally, one last trend is worth noting. We have pointed to the growth of small industries, particularly in clothing and textiles, and their implications for women's employment. The 1980s, however, have seen a more generalised growth in small business. It is impossible to say to what extent these businesses have contributed to women's employment, although if trends in industry are an indication, it may be significant. Figures from the SBDC, however, suggest that very few women are owners of small businesses. Out of a total of R215 588 713 lent in Natal since its inception, only 6% went to female clients. Women's loans are probably smaller on average than men's, so the figure of actual clients may be higher. Nevertheless, the figures are probably a reasonable reflection of women's position (as owners) in small business, relative to the position of men.

1.2.2 Employment by Occupation:

Women and men are not only concentrated in different sectors in the economy, but within sectors, women also exhibit different occupational patterns from men. Most women tend to be employed in a narrow range of occupations, where certain jobs, such as nursing, clerical and secretarial work, and domestic work, are highly "feminised".

There are also important occupational differences amongst women. African women, for example, are under-represented in white-collar jobs, and over-represented in less secure, low status jobs in the service sector. White women, and to a lesser extent, Indian women, tend to have access to relatively higher status jobs in clerical and sales, and professional occupations.

The extent of occupational segregation between men and women, and amongst women, can be inferred from table 9 which identifies the distribution of the workforce between occupational categories by gender and race. Table 9 illustrates that women are located primarily in service occupations. in clerical and sales, and in professional and technical employment:

• Women form just over two thirds of the labour force in region E, yet accounted for approximately two thirds (65.2%) of all workers in service occupations, 47.7% of all those in clerical and sales work and 44.5% of all workers in professional and technical jobs. Of all artisans and apprentices, however, only 5.4% were women, and less than one fifth (18.4%) of all managers and people in executive and administrative positions were women.

The proportion of women in these occupational categories, however, differs by race, and largely parallels race differences in sectoral employment:

• White women are over-represented in white-collar occupations. Although white women accounted for only 10.7% of all economically active women in 1991, they formed 62.8% of women working in managerial and administrative positions, 38.7% of women in clerical and sales, and 23.6% of women employed in professional and technical occupations. Very few women in service occupations were white (2.7%) and only 1.1% of all women working on farms were white.

• Although coloured women formed 1.9% of the female workforce, they accounted for 4.5% of all women in clerical and sales occupations, and 4.2% of all women employed as artisans and apprentices.

• Indian women accounted for 9% of all economically active women, but formed one fifth (20.2%) of all women in clerical and sales, 16.9% of women in managerial and administrative positions, and 23.7% of all women involved in production and related occupations.

• African women are over-represented in farming and service occupations. 98.5% of women classified in farming and related employment were African women, most of whom are likely to be farm workers (Maconachie, 1989). Of all women in service occupations, 94.1% were African women, and as discussed below, most of these women were employed as domestic workers. African women are underrepresented in all higher status occupations and formed only 17% of all women in managerial and related positions.

• Of all workers whose occupation was not specified, or not classified, more than half (53.9%) were women. 91.2% of these women were African.

Within these broad occupational categories, women are further crowded into certain kinds of jobs. Unfortunately, figures for women's involvement in specific occupations are not disaggregated for the various regions, but are available only for South Africa as a whole. These data, however, are likely to be similar in region E and give some indication of the extent of crowding in specific occupations.

The majority of women, listed in the professional and technical occupational category, work in health professions and in education. In 1991, in South Africa, "medical, dental and related health services" occupations accounted for 19.6% of all professional and technical workers. Approximately three quarters (74.5%) of all health workers were women, and 80.9% of these women were involved specifically in nursing.

"Education and related" occupations accounted for more than two fifths (43.5%) of all professional and technical occupations. Of all workers involved in education, 60.1% were women, and of these, 93.1% were teachers. Only 1% of women in education were in high-status positions such as rectors, principals or inspectors, compared to 4% of men. 3.3% of women in education were involved in tertiary education, whereas 8.2% of men were employed in tertiary institutions as lecturers or professors.

More than a quarter (26.6%) of all economically active women in South Africa were employed in service occupations. The majority of these women were domestic workers. In 1991, domestic work accounted for 52.5% of all service occupations. Of all domestic workers, 95.5% were women, and of these 90.1% were African women. The significance of these statistics cannot be overstated. Domestic workers are amongst the most disadvantaged of female workers. Historically they have been excluded from labour legislation such as the Basic Conditions of Employment Act, the Wage Act, the Labour Relations Act, Machinery and Occupational Safety and the Workmen's Compensation Act. As a result, domestic workers have not been legally entitled to annual leave, sick leave' (paid) maternity leave, unemployment insurance, a minimum wage, and a limit on maximum working hours.

In 1991, the Minister of Manpower made a commitment to domestic worker legislation, and specifically, to the extension of the Basic Conditions of Employment Act to domestic work. In terms of this Act, domestic workers will be entitled to a 46 hour working week, overtime pay, one month's maternity leave and two weeks paid leave per year. Although these rights represent a major breakthrough for domestic workers, they are likely to be very difficult to enforce in "private" households. Furthermore, causal domestic workers, employed as weekly chars for example, are most likely to be excluded from these benefits.

Analysis:

Several studies have noted that the labour market in South Africa is segmented by race, where workers of different race groups have been "channeled into specific segments according to institutionally determined barriers" (May, 1990, quoting Ryan, 1981). This study of region E suggests that the labour market is also segmented by gender, where the sex-typing of activities has limited women employment to specific sectors, and within these sectors, to certain kinds of jobs. The majority of women work in "secondary sector", low-status jobs, characterised by poor working conditions, lack of secure employment and promotion prospects, and, as will be discussed, low wages.

Occupational barriers between men and women in the labour market have been created through pre-market and in-market segmentation. Pre-market segmentation reflects the ways in which people's entry into the labour market is constrained. A major form of pre-market segmentation has been differential access to education along spatial, racial and gender lines. Spatially, access to education is better in urban than rural areas, and is particularly poor in rural Natal. Ardington (1991) argues that as many as a third of African children of schoolgoing age are not at school. In rural areas, lack of proper planning of facilities for Africans has meant that the location and provision of services tends to be ad hoc and incomplete. In rural KwaZulu, schools are generally initiated by local communities who apply for subsidies from the KwaZulu government. In rural Natal, Africans generally have to rely on farm schools built by farmers, very few of which are high schools. According to Ardington, 71% of pupils in schools in rural Natal are below standard 2, and 96% are below standard 5. The equivalent figures for urban areas are 43% and 71%, and 57% and 81% in rural KwaZulu. Generally, high school education is a rare commodity in rural areas. In effect, particularly in rural Natal, women and men are trapped by their isolation and the limited set of opportunities to allow them to get skills which can be used elsewhere.

Both boys and girls are affected. Hofmeyer (1993) suggests that there is basic parity between the educational achievements of men and women. Other studies, however, identify slightly different trends. Jacobs (1990), for example, cites a larger proportion of African secondary school-goers as being female (54%). Whether African women have had equal education to that of African men, or whether it has been slightly better, is a question for further research. Notwithstanding, African women do not occupy an equal or better position than African men in the labour market.

Education between females and males has also been sex-typed, both at school and in tertiatry institutions. Of all students registered for engineering degrees at Natal University, for example, only 7.8% were women. Nearly two thirds (62.7%) of all students registered for education degrees or diplomas, however, were women.

Another important factor influencing access to the labour market, and as implied earlier, is spatial location. In their study of the labour force in KwaZulu, May and Rankin (1990) highlighted the ways in which gender, location and economic activity are interlinked. They illustrated that economic opportunities differ by area, where employment opportunities are greatest in urban, and lowest in rural, areas. Access to urban areas, however, differs by race and gender. The creation of bantustans and blocked urbanisation constrained the nature of African people s involvement in urban employment. Within these constraints, however, African women historically have been less mobile than African men, and have had greater difficulty in migrating to urban areas. In recent years, the migration of African women to the metropole and informal settlements has increased (see Todes and Smit, 1993), reflecting the increased employment of African women in urban areas.

In-market segmentation refers to the ways in which people s access to employment opportunities in the labour market are restricted. These barriers result in forms of job discrimination, where certain workers are denied access to certain sectors, and within these sectors, to higher-status and higher-paying occupations. Restricted mobility between jobs is reinforced, to some extent, by hiring practices in industries and individual firms. Once a job is vacant, employers decide who to employ and at what level a person should be employed. Management further decides on the rate of promotion and on employees' access to on-the-job training.

The major barrier to women's employment in South Africa, and internationally, has been created through assumptions around gendered roles and abilities. The low level of women in managerial and high-status jobs, for example, partly reflects the view amongst employers that women are bad investments because of their family responsibilities. In a survey conducted by the Executive Women's Club of Financial Mail's top 100 companies in the country, more than half of the human resource managers polled did not want to promote women to managerial positions on the grounds that women's family commitments compromised their careers. White women, in particular, who tend to have more access than other women to higher status jobs, confront a "glass ceiling" in occupations, where their employment prospects are more limited than white men in similar kinds of jobs.

Women's domestic responsibilities have also influenced the nature of paid work which is available to them. Many jobs performed by women are extensions of work undertaken in the household unit. Much of women's employment in manufacturing, for example, has been in clothing and textiles industries. Similarly, as Taylor (1985) observes: "jobs in the services (as [secretaries], cleaners, filing clerks, canteen workers, [nurses and teachers] dovetail so neatly with women's domestic roles ... that they have quickly become seen as women's jobs in most countries" (1985, 33).

Job discrimination and occupational barriers to women's employment reflect, and serve to perpetuate, the sexual division of labour. In an attempt to overcome these barriers to women's employment, some countries (Australia, for example) have ratified "International Labour Organisation (ILO) Convention 158 on Workers with Family Responsibilities". ILO 156 aims to ensure that people with family responsibilities who work or who wish to work do not suffer discrimination as a result of their responsibilities to their families. The implementation of ILO 158 has included workers' access to flexible working hours, the availability of permanent part-time work, rostered days off, and extensions to parental leave provisions. Although the adoption of these measures in South Africa would improve women's access to paid employment, they would not necessarily challenge institutional constraints, such as the kinds of jobs that women are employed in, nor more generally, the view that women, and not men, have family responsibilities. Furthermore, they would not recognise the importance of reproductive activities for the functioning of the economy.

1.3 Employment opportunities, wages and access to income

There are two main sources of income for individuals: income that is earned directly in the form of wages or indirectly, through pensions and other transfer payments; and income that is "shared", where individuals that are part of a household rely on other household members for income. This section focuses largely on income that is earned directly and examines the nature of, and possible reasons for, wage differentials between men and women, and amongst women, in region E. Unfortunately, specific statistics on wage rates are not available in the census reports Rather, census data record levels of income that are received and which could include other sources of income aside from wages (such as pensions and remittances). In the absence of alternative statistics, however, income levels will be used as a proxy for wages where necessary.

Before 1981, many industries in South Africa paid separate wage rates for men and women. The maximum wage rate for women workers was set at approximately 80% of the equivalent wage rate for men (Budlender, 1984). In 1981, the Labour Relations Act was amended, and discrimination on the basis of sex was made illegal. This legislative change, however, has not implied that the wage differential between men and women has been eliminated. Firstly, the amended Labour Relations Act does not apply to employment in the public sector, where a large proportion of women are employed.

Secondly, it has been suggested that many employers have renamed "female grades in which sex is not specified but where the vast majority of employees are females" (Pillay, 1985, 28). By changing the job title of jobs where most workers are women, employers are able to bypass the law and legally pay a lower wage for work undertaken by women.

Thirdly, legal prohibitions on wage discrimination have not addressed, and indeed cannot address, the nature and implications of occupational segregation between men and women. The differential access between men and women to employment opportunities has implied differential access to rates of remuneration. Many of the specific jobs undertaken mostly by women are seen as low-status jobs, and accordingly, are amongst the lowest paying of all occupations.

Furthermore, within occupations, men and women often do not have equal access to non-wage benefits, such as medical aid, housing subsidies, pension benefits, and paid leave. Simply comparing wage rates, therefore, may ignore these hidden benefits and costs and the ways in which these are distributed between men and women. Although these benefits and costs are not amenable to immediate quantification, they are an additional factor to be considered when examining gender differences in access to income.

Levels of Income between Women and Hen:

Census figures for 1991 indicate that there are substantial differences between the levels of income received by women and men. A larger proportion of women than men receive no income at all, or an income of less than R3000 a year (R250 a month). In all subsequent income categories, however, the situation is reversed and proportionately more men than women receive an annual income of R3000 or more. Women are therefore over-represented in the very low income ranges. Although the percentage of all men and women earning high incomes is low, there are significantly more men in this income range than women.

• 61.4% of all women and 41.7% of all men over 14 years of age received no annual income in 1991. Although the reading of these statistics must recognise that a smaller percentage of women than men are economically active, it should also be noted that all those women who are classified as "not economically active" are not necessarily married and supported by income-earning spouses.

• 18.7% of women received an annual income of less than R3000, or a monthly income of less than R250. 15.9% of all men were classified in this income category.

• 10.1% of women, and 18.7% of men received an annual income of between R3000 and R10 000 a year (representing a monthly income of R833.33 or less).

• 7.7% of women received an annual income of between R10 000 and R50 000, whereas 18.4% of men were classified in this income range.

• 0.4% of all women, and 3.3% of all men received annual incomes of R50 000 and more.

• 2% of women, and 1.7% of men were listed in the "unspecified" category.

Income Levels and Occupations:

A primary factor accounting for income differences between men and women is differential access to, and employment in, the labour market. Women and men are employed in jobs which have different employment prospects, which are valued differently in the market, and which are graded differently by employers. Even within similar occupational categories, however, women earn less than men, reflecting wage discrimination (particularly in public sector employment) and unequal access to higher-status occupations and promotion within occupational categories.

The 1991 census indicated that two occupations in which women predominated over men were education and service employment. Statistics which correlated levels of income with occupations for gender and race were not available for region E. It is possible to use data for South Africa as a whole, however, to illustrate income distributions between men and women within these occupational categories.

Although more women than men were employed in education, women were over-represented in the low income categories:

• 5.6% of all women, and 4.4% of all men, received an income of between R1 and R3000 a year.

• 12.7% of all women, and only 6.6% of all men in education and related occupations received an income of between R3000 and R10 000

• The majority of women (59%) in education earned an income of between R10 000 and R30 000 a year, or a monthly income of R2500 or less. 47.0% of all men in education were classified in this income category.

• After R30 000, there are greater percentages of men at each income level than there are women. 2.5% of women, for example, received an annual income of between R50 000 and R69 999 compared to 10.6% of all men. Of all men and women in this income category, nearly three quarters (73.9%) were men and only approximately one quarter (26.1%) were women.

• 0.5% of all women and 4.2% of all men received an annual income of between R70 000 and R99 999. More than four fifths (83.8%) of all men and women in this income category were men.

• Although 0.1% of women and 0.1% of men were reflected in the highest income category of R300 000 and more, 62.7% of all those receiving this income were men.

The skewed distribution of income between men and women in education and related occupations can be accounted for in two ways: firstly, within education, women are in lower-paying jobs than men, and absolutely and relatively fewer women are employed in senior positions; and secondly, within similar or the same Job categories' women on average receive less income than men. It is not possible, however, to isolate these two factors from the 1991 census figures for income.

The distribution of income between men and women in the service occupations reflects similar characteristics:

• More than half (57.2%) of all women in the service sector received an annual income of between R1 and R2999, probably reflecting the high percentage of African women employed as domestic workers and farm labourers. Only 17.5% of men fell into this category, a percentage which would also include men employed as farm labourers, as well as those working as gardeners.

• Nearly two fifths (38.1%) of all men employed in the service sector fell into the middle income ranges (R10 000 to R49 999), compared to only 7.2% of all women.

• Although women outnumber men by 2 to 1 in the service and related sector, 85% of all workers who received an income of between R50 000 and R69 999 were men, representing 1.3% of all men employed in service occupations.

Income Distribution Amongst Women:

Tables 10, 11 and 12 indicate skewed distributions of income between men and women. When the income levels of women are examined more closely in table 13, however, significant differences amongst women can be identified, corresponding (at least in part) to racial differences in access to employment, and in kinds of employment.

Income distribution in region E by gender and race :

————————————————————————————————

Annual Income % Who Were:

(Women) White Coloured Asian Black

————————————————————————————————

R0.000 5.2 1.1 10.3% 83.4%

R1 - 999 5.1 1.2 5.8% 87.9%

R1 000-2 999 3.0 1.1 5.9% 90.0%

R3 000-4 999 10.4 1.7 9.6% 78.3%

R5 000-6 999 12.3 2.7 18.1% 66.8%

R7 000-9 999 16.2 4.5 27.0% 52.3%

R10 000-29 999 43.6 3.3 14.9% 38.2%

R30 000-49 999 67.8 1.9 12.5% 17.8%

R50 000-69 999 66.8 0.9 8.5% 23.8%

R70 000-99 999 61.8 0.9 8.7% 28.6%

R100 000-299 999 54.8 1.2 7.2% 36.7%

>R300 000 46.4 1.6 10.3% 41.7%

————————————————————————————————

Source: CSS report: 03-01-10 (1991)

Note: Zero income category has been adjusted, and all those under 15 years of age have been "removed".

Although African women form a larger percentage of all economically active women, they were over-represented in the low income categories:

• Of all women receiving no income, 83.4% were African women, reflecting the large proportion of African women involved in subsistence farming, who are unemployed, or who may have very irregular employment. Indian women accounted for 10.3% of all women receiving no income, and white women, 5.2%. A low percentage of all women in this category were coloured women (1.1%), corresponding to the relatively high rate of participation of coloured women in the economy.

• 87.9% of all women who received less than R1000 a year, and 90.0% of those who received between R1000 and R3000, were African women. Although these figures were not correlated specifically with occupations, it can be suspected that many of these African women were engaged in domestic work and farm employment.

The majority of women who received incomes in excess of R10 000 a year were white women, reflecting their greater access to relatively higher-paying, "white-collar" jobs (particularly in professional and clerical/sales occupations):

• Of all women receiving between R30 000 and R49 999 a year, 67.8% were white. White women also accounted for 66.8% of all women listed in the R50 000 to R69 999 category, and 61.8% of women in the R70 000 to R99 999 range.

Summary and Analysis:

This study has suggested that income differences between men and women, and amongst women, must be understood in the context of complex socioeconomic processes. Different levels and types of education cannot be discounted, but a human capital explanation does not adequately account for these earnings differentials. In fact, women's education levels may be somewhat better than those of men's. Rather, there is evidence of both wage and job discrimination in region E's labour market.

Wage discrimination results in different rates of pay for workers in similar kinds of work. Job discrimination, the corollary to segmented labour markets, reflects those processes whereby men and women have had differential access to employment, and therefore, to income-earning opportunities. The sex-typing of activities has restricted labour mobility between and within sectors. Women have also been seen as bad investments for on-the-job training, restricting their access to secure, upwardly-mobile forms of employment. The intersection between gender and race .discrimination has implied that African women are the most disadvantaged of all workers. Much of the work which African women perform is not unionised or falls outside the grading system in industry, implying that these workers are not covered by protective labour legislation and usually do not have access to non-wage benefits.

The existence of wage and job discrimination between races reflects apartheid policies and ideologies, which have enforced an economic, social and spatial segregation of the labour force as capitalism has developed in South Africa. Policies such as influx control, and job reservation in industry, and the control of unionisation, have undermined African people's access, in particular, to income-earning opportunities.

Discrimination between men and women ultimately reflects assumptions about the sexual division of labour and the kinds of work which men and women should perform. This sexual division of labour, in turn, assumes that men and women co-exist in households which are united and in which the allocation and sharing of resources are guaranteed by emotional commitments and altruistic behaviour. Under these circumstances, men (as the heads of households) "need" to earn more than women so that they can provide the financial support for their dependents.

Although the relations between men and women within households have not been adequately explored empirically, studies which have been conducted challenge the notion of a united household where resources are all pooled (see for example, Lund 1993). Rather, they point to living arrangements where women are disadvantaged in bargaining because of their lower levels of income, and where many decisions over resource allocation are made by men. Rural households have also been split up by the migrant labour system, where many women do not know what their husbands earn or even where they work.

Many households in South Africa are not male-headed, and women increasingly are supporting and maintaining their families. In 1980, Simkins (1986) estimated that amongst married couples, 59% of rural households in the bantustan areas were female-headed, as were between 36% and 47% of urban households in these areas. These figures may be attributed partly to the effects of influx control, male migrancy and the absence of family accommodation and social services in urban areas, policies which divided African families. Although influx control was scrapped in 1986, and African women legally are now permitted to live in "common" areas and to join their partners in search of work, many women still live without their husbands.

Several studies have also documented the increase in the number of women who choose not to marry, and who decide to raise families without forming permanent attachments with the fathers of their children, or with other male partners (see for example, Hellman 1974, van der Vliet, 1982, Jaffee, 1988, Preston-Whyte and Zondi, 1989). May s (1993a) figures for the region suggest that between 28.5% and 40% of all households are female headed, with the proportion highest in pert-urban areas, and lowest in rural areas. An important factor accounting for this trend concerns women s control over their resources. Studies have documented that many African women, for example, are challenging the view that marriage will ensure women s long-term security. Hellman (1974) and van der Vliet (1982) suggest that resistance to traditional family structures often comes from older African women who have been widowed, divorced or abandoned, and who have managed to raise their children on their own. In their study of teenage mothers in Kwa-Mashu and the greater Durban area, Preston-Whyte and Zondi also noted perceptions amongst some African women that female-headed households were likely to be more prosperous largely because they would not be subject to the drain on resources imposed by husbands who go about drinking your money and spending it on other women" (1989, 4). Budlender (1991) suggests further that men are more opposed to contraception than women. Consequently, single women may have greater control over the size of their family than married women.

Although some female-headed households have achieved a measure of economic security and success (Jaffe, 1988), on average, they tend to be in a vulnerable economic position. May's study (1993a), for example, identifies "a very strong gender bias" in the incomes of households, and points particularly to the high percentage of female-headed households which receive no wage or remittance payments, but which are dependent on welfare transfers only. With few employment opportunities in rural areas, female headed households in these areas tend to be the most disadvantaged. But even in urban areas, female-headed households must survive on the low levels of income earned or received by women. The proportion of welfare dependent households, the majority of which are female-headed, is surprisingly high in urban areas. Although the figures are much higher in rural areas, May's figures suggest that the pattern of dependence on grandmothers' pensions is not a purely rural one, as is sometimes assumed.

Female-headed households are most likely to depend on non-wage earnings. The percentage of female-headed households in May's sample which mix wages with informal and other income sources ranges from 30.7% in rural areas to 53.3% in metropolitan areas. Conversly, households dependent entirely on wages are most likely to be headed by men. In the case of the high-earners, the proportion of women-headed households drops to below 20% in most areas.

In the face of these realities, arguments which suggest that there are complementary roles for women and men, that these roles represent the optimum allocation of human resources (Becker, 1980), and that men therefore need to earn more money than women, lack persuasion and substance. Rather, these views serve to impose traditional assumptions on complex gender relations, and to perpetuate inequities between men and women in access to income-earning opportunities, and to other basic resources.

WOMEN IN THE INFORMAL SECTOR

The informal sector is a major area of employment for African women in the region. As noted earlier, available statistics do not adequately recognise and describe the informal sector, although a number of surveys provide valuable insights. A strength of some of the more qualitative studies is that they highlight the interrelations between women's position in the workplace and in the home, and how gender relations affect women's access to income-earning opportunities. Given that more general (and enumerative) arguments about the informal sector have been dealt with in the previous analysis, this section will focus largely on these gendered aspects of women's informal work.

2.1 The informal sector and income-earning opportunities:

Women form a significant part of the informal economy of Natal. If the census occupational category "undefined" is an indication of the informal sector, over half are women, and the largest proportion of African women work in it. Informal work is the sole, or main source of income for some households - particularly those headed by women. Studies suggest that women in female headed households are more likely to be employed in the informal than the formal sector (Naidoo, 1993). In May's (1993a) study, some 53% of metropolitan households classified as "mixed income wages secondary" are female headed. Sneddon (1990) shows that some 75% of women clothing sellers in Warwick Triangle are sole earners in women headed households.

The informal sector has experienced enormous growth over the past decade. Both "push" and "pull" factors have been important (May, 1993a). Bonnin et al (1991) suggest that deregulation, and the need to extend survival strategies in conditions of slow formal employment growth, have been important factors. Interestingly, May (1993b) also shows a marked growth of earnings from the informal sector in rural areas.

The growth of the informal sector has, at one level, extended economic opportunities for women in the region. As a number of studies have shown, however, it is not an unqualified good. There are substantial differences in opportunities open to men and women, and between urban and rural areas. There are complex constraints on the extent to which women can experience the informal sector as a "maximization", as opposed to as a "survival" strategy. Studies on the informal sector highlight the impact of the responsibilities for work in the home, and of gender relations within the household, on women's position in the informal sector.

Most studies on the informal sector point to the fact that in general, women earn less than men in the informal sector. May (1993a) shows that the average earning for women is around R330pm, compared to average monthly income of R865 for men. The subordinate position of women in the labour force, therefore, is replicated within the informal sector, with men dominant in the more profitable transport and production oriented service work, and women more prevalent in trade and homecraft related informal work (May, 1993a). Women are generally confined to activities close to, and often arising out of domestic work (see Bozzoli, 1991). By contrast, men tend to capitalise on non-traditional skills and combine old with new skills (Preston-Why/e and Nene, 1992). Even where both women and men occupy the same sectors, for example in hawking, studies point out that, in general, men earn more than women (Pentz, 1993; Naidoo, 1993).

Women tend to be concentrated in four main areas: prostitution, service work, petty trading and home-based piece work. Most of this work will lead to fairly low incomes, although prostitution is exceptional. Prostitution offers the highest income to women in the informal sector much more in many cases than can be earned by women in the formal sector (see Posel, 1993). Compared to other types of small business, the startup costs are also extremely low, with no need for capital investment. Preston-Whyte (1992), for example notes the case of a woman who worked as a domestic worker so that she could get accommodation close to opportunities for prostitution where she earned more in a week than in a month of domestic work. Nonetheless, there are enormous differences in income earned between different groups of prostitutes, with race being a major dividing line (Poser, 1993). Urban prostitutes generally do much better than those in rural areas. Further, prostitutes prepared to have unprotected sex earn much more than those unwilling to risk AIDS.

2.2 Deregulation and gender relations:

Much of women's informal work is partially or wholly illegal, and involves constant encounters with police, and police harassment. Although deregulation may alleviate some of these problems, deregulation is not a simple good for women. Even where informal trading is "deregulated" in Warwick Avenue Triangle, for example, harassment is a constant part of life as many traders are unaware of the procedures for acquiring permits and do not have them. Some are also engaged in occupations which are considered unhygienic or illegal.

Naidoo (1993) argues that while deregulation has eased police harassment, it has also facilitated the entry of formal Indian male traders using local women as a front paying them low wages. These traders further have brought in manufactured goods which have displaced some of women's previous activities of producing goods for sale. This is consistent with Friedman and Hambridge's (1992) warning that with deregulation, activities could become more profitable, encouraging men to take over. More generally, however, men tend to move into areas of informal activity when they become lucrative as Preston-Whyte and Nene's (1992) study of KwaMzimela shows.

Preston-Whyte and Nene (1992) suggest that the more successful women in the rural informal sector depend on male cooperation to succeed, with a trade-off of continued subordination. In the case of rural areas, access to the experience of, and skills learnt in, urban areas is critical to success. Women are far less likely to have been to urban areas, or to have acquired these skills than men Where women had done so, they could diversify into new areas and tended to earn a better income. Where women earned more through working with their husbands, however, they were not necessarily in a position to control the money coming in, as men largely determined expenditure.

2.3 Domestic work, childcare responsibilities and informal work:

Although the informal sector offers a rather limited range of opportunities for women, it is better, in some cases, than opportunities previously available. In addition, it allows women a flexibility (in, for example, negotiating between the demands of duties at home and work) which formal employment does not.

Preston-Whyte s study (1992) supports these suggestions when she notes that since the 1980s, there has been a shift out of domestic work into the informal sector as the latter has become less risky, and women are no longer dependent on domestic worker accomodation for access to the city.

Naidoo (1993) argues that the women she interviewed felt that working in the informal sector allowed them more independence and power than did working in the formal sector. An attractive aspect of the informal sector is that women can bring their children to work with then if alternative care is unavailable (although this was generally seen as an undesirable alternative). They can stay at home when necessary, and can get money whenever they need it. Women therefore can move in and out of the informal economy, often as the need arises. This flexibility is important to older women who may have to stay at home due to illness. The informal sector is not preferred by young women, however, who would, if possible, choose better jobs and wages.

The informal sector therefore offers women more flexibility than does employment in the formal economy, and allows some compromise between the tensions of work in the home and work for money. But responsibilities for domestic work and childcare still proscribe their performance in the informal sector.

Naidoo s (1993) work highlights the closer interrelation of conditions for women in the informal sector and their responsibilities in the home.

In her study of Warwick Triangle, she observed that many women were unable to get access to the prime spots for trading as they had to cook breakfast in the morning - a constraint which men did not face. In general, responsibility for housework limited the time women could devote to their informal work, yet long working hours are essential to profitable operations, particularly in hawking. Most of the women interviewed said that if there was another woman in the household who helped them with domestic work they would be able to devote more time to trading.

Naidoo s study therefore showed that the division of domestic work between women in the household influenced the position of women in the informal sector. The extended family structure offered some women, who were living with kin, greater opportunities to work outside of the home. Many women in this situation wondered how they would cope if they had not had such help. This is not always the case, however. Older women generally were freer from domestic tasks than younger women, often relying on daughters to do domestic work while they went off to trade. Furthermore, an extended family structure need not help to ease domestic work, but rather, can imply additional home responsibilities, and less scope for informal activities.

Women who do have to take responsibility for childcare, and who cannot rely on kin to assist or provide financial support, have to devise other means to cope. Some trade intermittently. Others stop work when children are small, but this is not an option for many women in the informal sector who need income on a fairly regular basis to support their families. Long hours in the informal sector are also very burdensome for women with children. The inadequacy and inaccessibility of child care services for women in townships and informal settlements has meant that women have had to create a range of informal child care networks. At work, they rely on other traders in neighbouring stalls while serving a customer. In Naidoo s (1993) study none of the women had continuous and regular help with the children and most did not earn enough money for a childminder, and so were forced to rely on complex arrangements based on mutual exchange of services.

2.4 Urban - rural differences in income-earning opportunities:

Drawing largely on May and Rankin's study (1990), the analysis in the previous section noted that rural women earn far lower incomes than women in urban areas, and that there are significantly fewer employment opportunities for rural women than men. Work in the informal sector mirrors this spatial divide. Studies suggest that opportunities for informal sector work - especially in areas such as trading and services where women predominate - are far higher in urban areas and in mayor cities than in rural areas, presumably due to agglomeration economies.

In effect, as Preston-Whyte and Nene (1992) and MacIntosh (1987) show, the spatial marginality of rural areas, the low incomes and the relatively dispersed populations mitigate against a significant growth of the informal sector. The more profitable parts of the rural informal sector rely on large external markets (for example, handicrafts), trade along major routes, or on those times when incomes are larger (particularly when pensions have been received).

Preston-Whyte and Nene (1984) suggest that social constraints also limit rural women s prospects in the informal sector. Young women require their husband s permission to sell beyond the home, and social custom proscribes what women can and should do.

The formation of producer groups is one way in which rural women engaged in various types of informal production (especially clothing and crafts) have attempted to overcome spatial marginality, and their lack of capital and tools (McIntosh, 1992). These groups, however, face various gendered constraints, which render them marginal and intermittent. A critical factor influencing the performance of producer groups is the availability of cash for investment - which itself depends on whether women have the independence to allocate money to group activities. Further. the activities of producer groups compete with housework, and husbands and mothers-in-law may limit the involvement of young women in these undertakings. Many of these groups therefore tend to be dominated by older women. Groups managed by outside people overcome some of these constraints as women can work at home, and the groups do not depend on member contributions. But this kind of Organisation runs the risk of promoting rural women's dependence on outside structures and of undermining any control over production which they may otherwise have been able to realize.

WOMEN IN THE RURAL ECONOMY REGION

The notorious role that economics statistics play in making women s work as well as their lives invisible has been referred to in earlier parts of this study. In considering women in the rural economy of region E we confront similar problems. Assumptions about men being breadwinners, and about households operating as undifferentiated units, plague our understanding of social and economic relationships in rural areas as well as in urban or informal settlements. Aggregated statistics do not tell us anything about gender relations within the household, and about gender differentiated access to resources or time. It is virtually nigh impossible to get accurate macro information on the labour that is necessary to meet basic needs for daily survival, on the extent to which women are casual agricultural workers, on the extent of subsistence agriculture, or on women s involvement in smallholder agricultural production. Where such data does exist it is to be found in the small and independent (often qualitative) studies that have been conducted by individual researchers.

This section of the paper will consequently rely on existing research to highlight the particular situation of rural women with respect to subsistence agriculture, smallholder farming, commercial farming, and income generation oriented development projects. The focus on women s participation in agricultural production, however, must be contextualised at the outset. Many studies have noted that, partly as a result of the non-viability of agriculture, a large percentage of rural woman rely on pensions, remittances and disability grants (Ardington, 1988; CORD, 1990; Gandar and Bromberger, 1984; McIntosh, 1991; McIntosh and Friedman, 1989).

3.1 Subsistence agriculture

May’s (1993b) sample indicates that the average income from both the sale and the consumption of agricultural activities made up only 4.7% of household income and that this has declined by 1.9% since 1985. For households further than 200 km from Durban, agricultural income increases as a proportion of total income. In a few rare instances this income could increase to as much as 25% (with a man living permanently at home and actively involved in agriculture). This activity is obviously not the mainstay of rural households existence. Nevertheless, many women are involved in maximising what they can from the available land. Numerous development organisations, and the KwaZulu Department of Agriculture and Forestry (KDAF), for instance, advocate and support community gardens as part of community development strategies. The reasons given to support this choice tend to include: they improve nutrition; they are useful as dietary supplements and contribute to food security; they are well suited to a self-help philosophy and development; they are a good adjunct to regular field crops. Given the demographic realities in these areas, largely due to the historical legacy of predominantly male migrant labour, and women s historical gender role as food producers, it is not surprising that it is usually women who tend to belong to these gardens.

In these projects, a group of women are stimulated to individually grow vegetables on a small plot of ground (usually about 100 m2) primarily to provide nutrient supplements for daily diet (Carr, 1991). The time and labour involved is substantial especially when water has to be manually carried from the water point to the field. According to the KDAF, some 31 957 women participate in 1041 community gardens that are serviced by their extension officers (Brooks and Friedman, 1991). Production figures and income generated from these gardens were not available for all the sub-districts, but average figures for 1989 for the Mabedlana district point to half the total produce being sold. Individual women that have been interviewed who are fairly successful in this activity have expressed interest in expanding their production for sale but feel frustrated by the environmental constraints (lack of transport and marketing infrastructure). Cases where gardens have been organisationally linked with a view towards broader goals have facilitated interesting developments. The Thuthukani Consumer Cooperative in Mangusi was built up from a strong base of community gardens and currently services a few thousand members (Hulbert, 1991). Agricultural inputs are obtained by members at prices below the average retail price and the co-op serves as an important institutional structure in the area. Other services such as water development have recently begun to be facilitated through the coop structures.

Community gardens have been described as the backbone of subsistence agriculture in KwaZulu (DBSA, 1988), yet very little information is available. Although membership of these gardens is large, their output is so small in relation to the total agricultural output that little research has been done on them. In reality, community gardens are largely a catchnet for welfare purposes, serving the very poor households who have a desperate need for land (Lyne, 1991).

3.2 Smallholder farming:

van Zyl et. al. (1993) note that the production of sugar-cane is the most important agricultural enterprise in the region. They also indicate that there are approximately 350 000 ha of sugarcane farmed by large-scale quota holder farmers and 120 000 ha by small cane growers. Vaughan has done extensive research into the topic of small cane growers and estimates their growth from 30000 at the end of 1989 to 42 313 by August 1993 (Vaughan, 1993). The majority of these farmers are contract farmers which essentially means that they are bound to sell their harvests to the company with whom they are contractually related. Usually the growers receive credit, inputs, machinery and technical advice from the company. Responsibility for specific tasks across the spectrum of cultivation tends to vary. The company, however, often delegates some or all of the responsibility for production, but retains control of processing and marketing (Vaughan, 1993). The issues of contracts, of the number of small growers that are involved in the various cane schemes in region E and the potential benefits from the income are of considerable importance when considering women in the rural economy.

With regard to the contracts, a registered grower has to sign the contract with the company. "To sign a contract with a milling company one must have rights in land, and the legal status to enter into a contract. Their legal status and lack of direct rights to land are the factors which may prevent women from being producers in their own right" (Vaughan, 1993:19). Once the cane is harvested and sold, the income generated thereby is paid out directly to the registered grower regardless of who has been involved in the labour of producing the cane. Hence, when women are the de facto producers but not the legal registered growers, they may be deprived of the income derived from the cane they have produced (Vaughan, 1993).

This point is particularly highlighted when one recognises that 65% of the small grower producers are women (cited by the Small Grower Liaison Officer in Vaughan, 1993). Far more detailed research is necessary to measure the extent of dissatisfaction with this reality. Scant evidence, however, suggests that the women are not altogether content with the status quo. They have organised to get paid out for the task of weeding at convenient times when the registered men are absent. This is a small way in which they retain some control over the proceeds of their labour. In other interviews, women have complained that they work the fields but don't have the money and cannot therefore make big decisions like joint purchasing of tractors (Vaughan, 1993a).

The few women who are registered in their own names are generally widows who can inherit land. Legislation passed as recently as 1987 now allows women who are heads of households to acquire land. As long as a woman is unmarried and has dependents (either children or parents) she can now legally acquire land. At least this change in legislation acknowledges the existing marginalised status of many women headed households and the feminisation of poverty. However it does not tackle or challenge fundamental patriarchal power in any real way and still ensures that married women have no independent rights and are fully dependent on their husbands. In a situation with high rates of migrancy such as in rural areas of region E, this legal status quo is experienced as an economically debilitating factor for women. "Production levels are undermined in the prevailing situation where rights in land and in the crop are often vested in absent men, rather than in the women who are responsible for production" (Vaughan, 1992:435). While the issue of women's rights to land and their control over its product in areas of tribal tenure needs to be addressed at a national level, the solution is not simply in a private market in land. It must be remembered that "the preservation of traditional modes and patterns of land allocation has meant that people's existing rights are not undermined and threatened. Marginalised and rural people have been able to participate in production - most particularly women and children who have no source of income other than that which derives from the cultivation of cane on the fields to which they have usufructuary rights" (Vaughan, 1992:432).

Vaughan (1993) concludes that cane production has given many economically marginal women an opportunity to participate in production and to generate at least some income on land which otherwise might not have been utilised at all. Even though this income is far from adequate, it does provide an important supplement for those who do have alternative income sources, and a crucial mechanism of survival for those who are entirely dependent upon it. Elsewhere, Vaughan (1992) has described how income from cane in Kwazulu has been invested largely in education, housing, and consumer durables. Significantly it has enabled people to pay for secondary and tertiary education for their children and has facilitated a markedly improved quality of housing.

The timber industry has not yet developed its outreach into the small grower market to the extent that sugar has, but we can see a steadily growing interest in this direction. Sappi and Mondi combined are involved with approximately 4000 small growers (Cellier, 1993). Despite the differences in nature between the crops, there are certain similarities in dynamics between the two industries when it comes to contracts and small growers. As is the case in the sugar industry, the contracts are usually signed with the men even though women (usually their wives) are the de facto producers. In some instances, women have also organised with the foresters to get what payouts they can when the men are absent

(Friedman, 1991). Less research has been done with respect to small timber growers than sugar and the length of crop rotation for timber is longer than for cane. Consequently, it is difficult, at this stage, to know much about what happens after harvesting.

Part of the reason that the contract farming model is attractive to small growers is that many of the infrastructural problems so dominant in rural areas (ea. access to transport, marketing, credit) is addressed by the company s involvement. The relationship does tend to bind growers to particular contracts and companies, but in the face of few other options with such high degrees of support, this is a choice that offers some relief to grinding poverty. The levels of organisation that are necessary to facilitate communication between growers and the company also offers some space for the social and human development of the grower membership. The development of such human and social capacity, however, requires careful nurturance and training.

The concerns raised above in relation to women and their control over production proceeds, migrancy, rights to land and power relations between men and women, are of critical importance in understanding the development potential of smallholder farming.

3.3 Commercial farming:

From January 1993 and May 1993 the Unemployed Insurance Act and the Basic Conditions of Employment Act (BCEA) were extended to apply to permanently employed (as opposed to casual or seasonal) farmworkers. In the past, there was no labour legislation to protect farmworkers, and the application of these acts has very important consequences for farmworkers. Farmworkers are now entitled to a restricted working week (48 hours/week), the right to paid leave, eight public holidays/year and sick leave. Women are entitled to three months maternity leave and workers are permitted to belong to trade unions. The only change for casual workers is that they are now limited to a certain number of hours per day. Thami Mohlaka, a lawyer working for the Farmworkers Research and Resources Project (FRRP) claims however that there is still inadequate protection for farmworkers against harassment from farmers and that there is absolutely no protection against unfair dismissals. Some people fear that the spate of recent retrenchments has in fact been influenced by the passing of the Act.

The conditions that determine most women farmdwellers' lives are unlikely to be affected by the acts. Firstly, this legislation does not apply to causal or seasonal farmworkers, and studies suggest that the majority of these farmworkers are women (Buijs, 1985; Marcus, 1985 and May, 1987). These women obviously would not benefit from the maternity leave provisions or any of the other major benefits mentioned above.

Secondly, several women who live with male farmworkers are generally expected to work as domestic workers on the farm by virtue of being the legally employed workers' wife. The acts do not cover these women. The other serious consequence for women and children who are dependent on male farmworkers is that in the case of eviction (with labour tenants) and retrenchments, they lose access to shelter and other basic resources (wood and water).

3.4 Income generation oriented development projects:

Given the severe lack of formal job opportunities for rural women, and the fact that more and more women are being forced to generate some form of income in order to survive, it is perhaps unsurprising to notice the proliferation of so-called income generation projects. The vast numbers of "self-help" groups that organise themselves around craft production, sewing, beadwork, knitting, basket making etc. but that earn very little, is testament to the levels of desperation that exist in the rural areas. These activities are virtually always marginal and supplementary (PrestonWhyle and Nene, 1984; McIntosh, 1991). Various development agencies in the form of State extension services, various non-government organisations, the church or feeding organisations, (ea. Operation Hunger) often resort to encouraging such "income generating projects" in the absence of any viable' visible alternatives.

All too often particular gender stereotypes which produce low returns, are employed in fostering these activities. Women are encouraged to build upon skills useful to, or acquired in, the domestic arena and are seldom stimulated to think beyond such stereotypes to get involved in activities that challenge the gender division of labour (ea. road building, welding etc.). In addition, structural support for the multiple other tasks that women have to perform for daily survival are seldom considered in the overall construction of such projects. This is one of the reasons why income generating asides will always remain so.

3 5 Concluding comments:

It is not surprising, given the limited income opportunities for rural women, that May s study (1993b) should suggest both a higher proportion of weekly migrants being women, and in a context of an overall reduction in migration, an increase in migration levels amongst women. In the absence of more precise data, these trends represent generalisations, but they raise important questions for subsequent research. For example, are most of the migrant women young, do they have older children or are they childless? Is an increasing proportion of remittances being sent to rural areas from women - either daughters or mothers?

Although no research has been done to elaborate upon the precise dynamics, studies also suggest that more men are returning to live in rural areas either due to unemployment, retrenchments or as refugees from violence. This trend is likely to put more pressure on women to generate income whether in male-headed or female headed households. It also means that the limited control women have over their domestic space in the absence of male migrants is further eroded with the men s return (Cellier, 1993).

SOCIAL REPRODUCTION AND ACCESS TO RESOURCES

Inequalities in the workplace and in the economy are mirrored and reinforced by inequalities and inadequacies in access to resources affecting social reproduction. Access to housing, physical and social infrastructure all affect women's economic position, and are affected by it, in complex and sometimes unseen ways.

Women's responsibility for reproduction, and its impact on their economic position, generally is not recognised in economic policy thinking. Reproductive tasks are critical to the survival of households and, more broadly, to society. Without the work of producing and raising children, cleaning homes, cooking, caring for the aged and sick etc., remunerative work would not be possible. Reproductive or domestic tasks generally are unpaid, however, and are not valued as work. Women engaged solely in unpaid reproductive tasks are classified as "not economically active", although the work that they do is not dissimilar to much of the work which is included in the formal economy. The invisibility of women's reproductive labour entrenches the assumption that the process of reproduction is a given which will simply "occur". The process of reproduction is naturalised, as is women's responsibility for this process (Poser, 1993).

Many feminist theorists emphasise the relationship between the position of women in the economy and their responsibility for reproduction. On the one hand, the discounting of women's unpaid labour in the home has adverse effects on women's participation in paid work. The assumption that women "naturally" are the reproducers affects the way women operate in the labour force, as our discussion of both the formal and informal economies has shown. On the other hand, women's position in the labour force, and their lower incomes, affects their access to those resources (such as housing, childcare, water, energy) which facilitate the process of social reproduction.

There are of course real differences between women in terms of race and class, with some women enjoying reasonable access, while others do not. But, all women, to a greater or lesser extent, are adversely affected by assumptions of gendered roles and abilities.

The discussion below will document women s access to resources of various kinds, and will highlight the ways in which access to these resources affects, and is affected by, their position in the economy. The general inadequacy of these resources for most African women can be seen as highly problematic for their participation in the economy. Furthermore, while the participation of women in remunerative activities has increased over the past decade, the conditions under which it has occurred has placed considerable stress on the women involved, and particularly where social networks and supports are not well developed. A major issue here is the conditions under which the generally unpaid work of domestic labour is occurring. The well documented inadequacies of physical and social infrastructure in the region have significant implications for the amount of work involved in domestic tasks.

A number of studies have demonstrated the inadequacies in, and disparities of, the distribution of physical and social infrastructure resources between race groups and areas in the region (see ea. Smith, 1993; Todes and Smit, 1993; Ardington, 1992; Hambridge and Krige, 1992; Stavrou, 1992; Krige, 1989 and 1990). In general terms, services are far worse for Africans than for other groups, urban and metropolitan townships are generally better off than informal settlements, while rural areas are usually worst off. Rural areas in Natal rather than KwaZulu are particularly poor in terms of access to education and health.

Inadequate infrastructure greatly increases the labour time involved in domestic work and in this sense constrains women s participation in paid work. In rural areas, labour time is not currently a major impediment on agricultural production (see May, 1993b) due to the myriad of other constraining factors. It would become so, however, if extensive rural development projects, focusing on increasing production, were initiated.

The impact of the additional work imposed by poor infrastructure, and by domestic labour more generally, is mediated within the household. The scope for mediation, however, is limited by gender constraints. It is almost always women who undertake the additional labour associated with poor infrastructure. Annecke (1992), for example, notes that while young boys assist with tasks such as carrying water, this activity generally stops by the time they are 12 years old.

In larger households, the division of labour between women allows at least some women in the household the opportunity to earn money. There may be costs associated with this division of labour, however. For example, because of their domestic responsibilities, daughters may be prevented from attending school or from giving much attention to their school work.

Smaller, nuclear households impose particular burdens on women who are working. In a study of households in different settlement types in Durban, Todes and Walker (1992) found that the classic 'double burden" on women operated most clearly in these instances. Similar points are made by Naidoo (1993).

4.1 Housing:

The process of urbanisation has significant potential benefits for women. As we have shown, and as demonstrated elsewhere (see May and Rankin, 1990; Todes and Smit, 1993), the economic opportunities open to women are significantly higher in urban than rural areas, and particularly in the metropolitan areas. While these differences are notable for men as well, the relationship is much stronger for women (May and Rankin, 1989). Women's access to urban areas, therefore, is important in affecting their access to employment and income. The growth of women's employment in urban areas, and especially in Durban, has been matched by increasing rates of urbanisation (see Todes and Smit, 1993). Some two-thirds of new entrants to Durban in the 1980s were women. Nonetheless, women headed households experience greater difficulties in gaining access to housing and land in urban areas than do those headed by men.

In the past, influx control and the rules governing access to housing limited the extent to which women could gain access to urban areas (see Todes and Walker, 1992). While influx control is gone and many of the rules have been altered in seemingly more gender neutral directions, women still face difficulties in acquiring accomodation.

Since very little new public housing has been built for rental over the past decade, access to public housing is largely the result of historical access. In the case of Africans, married men with dependents have received preference. Since 1979, women with dependents have been allowed to rent or own these houses. But since the available stock has not been extended, and people have tended to hold onto their houses, women have only really been able to acquire houses through inheritance. A study in Durban showed that the proportion of women owners and occupants in public housing outside of KwaZulu was higher than might be expected, primarily because the age of the housing has meant that women have "inherited" it, and these women have been protected by the "person plus dependent" ruling (Todes and Walker, 1992). The numbers of women benefitting in this way, however, are very small.

No studies have been done on white, coloured and Indian women's access to public housing, or on public housing in KwaZulu, although similar factors may operate.

Women's access to housing which is developed and financed through the private sector is much more limited. Although there is no explicit discrimination by financial institutions against women, their lower earning capacity affects their ability to acquire housing. Employer subsidies - particularly through the public sector - have been an important channel for gaining access to housing. Parnell (1991), however, shows that the calculation of the public sector employer subsidy effectively discriminates against Africans and married women. Todes and Walker's study of developer-built housing for Africans in Durban (1992) supports this proposition. Although African women form a large proportion of public sector workers, they found that only 15% of houses were owned by women. Subsidies available to buy houses have also operated in favour of the existing tenants, and most registered tenants have been men (see Parnell, 1991).

Given the constraints on African women's access to formal housing, it is perhaps not surprising that there is a higher proportion of women-headed households in informal settlements than in developer-built housing. Women's access to these settlements, however, is by no means "free". Cross et al (1992) suggest that informal mechanisms of control and rules may limit women's access to these areas. Ironically, the social norms, historically associated with the rules of access to public housing, are sometimes operative, with preference being given to couples with children over female or single headed households. Much more attention needs to be focused on where and how women acquire housing in informal areas and the informal rules which determine that access.

A number of schemes to develop sites and services in informal settlements have been initiated recently. Title to sites generally is part of the upgrading package. As previously, only "people with dependents" qualify for a subsidised site. While this qualification allows women household heads access to sites, married women are disadvantaged.

Although women put a great deal of effort into these development projects, they may well be disempowered as title to land goes to the man. Further, depending on their marital regime, they may not have a share of the property on divorce. Housing is often seen as a potential source of accumulation in housing policy. Granting title to men, however, may well give this potential to men, and deprive women of this accumulation source.

Over the past few years planners have emphasised the importance of the development of more "compact" city forms, and the development of land close to areas of employment for low-income people. For women, the economic logic underpinning these arguments is strong, as will be demonstrated below. A number of primarily women informal sector workers are beginning to live on the streets of central Durban, and there is pressure on the limited available accommodation for low-income women in central areas - the one women s hostel and domestic worker accommodation. Domestic worker accommodation, however, is beginning to be eroded by conversion to higher earning uses.

While women are urbanising and increasing their participation in paid work, they are doing so under difficult conditions. These conditions may limit the extent to which new migrants can see themselves as permanent urban settlers, and they may affect how women participate in the economy. The majority of women, and especially women-headed households live in informal settlements distant from places of work, and where conditions around domestic work are particularly onerous.

4.2 Water:

Possibly the most serious infrastructural deficit for women is the absence of an adequate water supply in most areas in the region. In Durban, some 52% of African households (corresponding to most informal settlements) live in areas characterised by a "minimum" service level, ie. less than one standpipe per 250 m (World Bank, 1993). Arrangements in these situations vary from fewer standpipes offering free water, to standpipes available for use at predetermined times, to water kiosks selling water at rates far exceeding rates applicable to reticulated water, to informal purchase and sale of water from neighbouring households, often at exorbitant rates, amongst other arrangements. According to Smith (1993), the situation in urban informal settlements is much worse than in metropolitan informal settlements, with many areas relying on streams, springs, rivers and tankers. There are no studies examining the gender implications of various water arrangements or of the gender implications of inadequacies in water supply. However, a number of points can be surmised about the implications of water supply deficits.

Firstly, it can be expected that a larger proportion of women headed households will be located in areas characterised by poor water supply than in areas where services are generally better. Secondly, since women headed households are on average poorer than maleheaded households, the higher cost of water in some cases (through kiosks or informal arrangements, for example) will generally impact on them more. Thirdly, compared to reticulated water, the labour time involved in the tasks which require water usage is very much greater. Fourthly, difficulty in access to water tends to reduce the level of water usage, with potentially health-threatening implications. Use of boreholes, streams, and rivers in urban informal settlements may increase susceptibility to disease. Again, it is women who tend to take responsibility for the sick, impacting on their participation in the labour force. Studies elsewhere (see Harris, 1989) have documented the significance of health in poverty and in precipitating households into conditions of poverty. Since women and women-headed households form a large proportion of the poverty grouping, they are particularly vulnerable in this regard.

While water supply is generally better in formal townships than informal areas, it is not unproblematic. Most (96%) metropolitan townships have reticulated water, but 40% of urban townships have only access to standpipes every 250 m or less (Smith, 1993). Poor management and under maintenance in KwaZulu townships, and increasing population pressure on them (see Todes and Smit, 1993) has meant that services are not necessarily fully functional, with implications for women s labour time. According to Smith (1993) some 70% of metropolitan townships and a few urban townships experience problems with their water supply. Unfortunately there are no local studies documenting its impact on labour time. White s (1991) detailed studies of similar conditions in Soweto, however, leave little doubt as to the additional labour burden it imposes.

Smith (1993) reports that rural areas have the worst water conditions. In the present drought, there is no water supply in a number of areas and tankers have had to be brought in. In effect, much of the day in rural areas is taken up by the tasks of fetching water, firewood and other household duties. In these areas. women are faced by a double bind: the work load involved in household work limits their involvement in other areas, coupled with enormous constraints on any potential economic activity arising from, inter alia, their spatial marginality.

Lack of adequate supplies of water may generate a number of marginal informal income earning opportunities which are generally taken up by women (such as water-carrying for example - see Annecke, 1993), but it is outweighed by their overall impact on the condition of Lou-income women. From the perspective of low income women, improvements in water are vital, but these improvements will have to ensure that women are not excluded by the unaffordability of solutions. Further, as elsewhere, the way improvements to water are made, and the inclusion of women into decision-making processes, will be critical.

4.3 Energy:

Over the past few years, programmes have been set in place to extend electricity to all areas in the region. In theory, the extension of electricity to all areas is in women s interests, since energy, like water is an integral part of domestic work, and alternative energy sources are in general more time-consuming, with the additional labour generally falling on women. For example, collecting wood in rural areas can consume as much as 4.5 hours per day (Smith, 1993), with women walking up to 8.3 km (DRA, 1992). Similarly, cooking with wood is far more time-consuming than with transitional fuels (such as paraffin), while electricity is far quicker than either. Per unit cost, electricity is cheaper than most other fuels (except where it is free, as may be the case with wood for example), and generally cleaner and quicker to use.

It is becoming apparent, however, that the introduction of electricity is not displacing the use of alternative energy sources, nor by implication, reducing labour time to the extent that might have been expected. Electricity usage is lower than expected, which may threaten the economics of the programme. The electricification drive is not necessarily benefiting women as much as might have been predicted on a superficial examination. A number of in-depth studies shed light on these issues, and raise questions about the extent to which electricity (at least in the present form) is a solution for women. These studies are exciting because they are amongst those few studies of infrastructure which go beyond a broad discussion of access, to looking at the social relations surrounding infrastructure - to a level where the influence of gender relations becomes apparent. A number of points can be drawn from these studies.

It is clear that gender power relations and control over decision-making within the household are critical. Data Research Africa (DRA, 1993c) conducted a study on households in four areas which had been electrified. An important insight of this study was their suggestion that heads of household (two thirds of whom are men) tend to make decisions about whether, and what kinds of electrical equipment, to buy. In many cases, more priority is given to buying items such as televisions and Hifi's than to purchasing equipment associated with alleviating the burden of household work, especially cooking. Similarly Gwagwa's (1993) study of Inanda shows that women generally put more money into cooking facilities than men, but have lower incomes to do so. Notably, low ownership of electrical equipment was seen as a major reason for low usage of electricity.

The above analysis suggests that female headed households would be more likely to prioritise investments which would lead to a higher usage of electricity for domestic work. Income is a major constraint on electricity usage, however, (especially as electricity usage tends to raise expenditure on all energy), and male headed households tend to be better off than female headed households. DRA's analysis of income and electricity usage shows that as income rises, so does the use of alternative fuels, but only up to the middle level of income. Thereafter, alternative fuels tend to become displaced by a move to sustained use of electricity. It is only at the highest income levels that electricity is used completely (DRA, 1993c). Nevertheless, the study also suggests that over time, electricity usage increases. Income will, however, remain a determining factor, and particularly, for very poor households.

Annecke's (1992) study sheds light on the relationship between energy usage and income in very low-income communities. She shows how the use of mainly paraffin and candles, usually bought on a daily basis, is intimately tied into household survival strategies. In Canaan, a very marginal informal settlement in the DFR where a large proportion of the women depend on garbage picking, the irregular and low incomes earned by women means that they frequently buy their fuel on a daily basis or every second day, depending on what they have earned. Apart from the occasional "feast", they prefer to allocate their earnings across a number of items, rather than buying more of one item. This means that they generally buy at high prices from local stores or spaza shops. Although buying in bulk is cheaper per unit, buying one candle or bottle of paraffin at a time means that they can more easily control their budget. For example, if they buy a packet of candles another member of the household may burn them all in one night.

Annecke continues that electrification may be a somewhat mixed blessing for women in such circumstances. Apart from the entry costs (R136 in the DFR and R30 elsewhere) it may be difficult for such households to budget in the way necessary for the use of cards, and it may weaken their attempts at budgeting.

Studies suggest complex relationships between the use of particular fuels for certain tasks and (generally gendered) social practices. The choice of fuels is not purely a question of cost or labour saving. For example, Annecke (1992) shows that the use of paraffin for cooking a meal - and its slowness relative to electricity - is preferred by the women of Canaan as it elevates their status in the household. It visibly demonstrates the effort they are putting into looking after the family. Sometimes bread is cooked as opposed to bought, and especially in hard times. Even though in fuel terms cooked bread is more expensive, women see it as special, filling, and as demonstrating their love of their family and children. Similarly, DRA's (1992) study of Ingwavuma shows how the use of wood for cooking ties into the daily routines of rural women. By contrast in an urban context, Annecke (1992) argues that women in Canaan, although extremely poor, would only use wood (although free, and generally available) for cooking in particularly bad times as it is messy extremely time-consuming and places time demands which do not fit with their work.

Studies which have argued that cooking will be one of the first tasks to rely on electricity' therefore, would seem to be incorrect. In addition to the reasons alluded to above, DRA (1992) suggests that traditional cooking methods also serve to heat up the home. They therefore tend to be preferred to electricity, despite the greater convenience of electricity. In addition, certain kinds of meals, which require very slow cooking, cannot not be made on electrical stoves or hot plates, a point which Annecke s study (1992) substantiates. Income also constrains the use of electricity in cooking - generally high utility, relatively low cost, items such as kettles and irons tend to be bought before more high cost items such as hot plates, and stoves (DRA, 1992).

These points raise questions about whether electricifcation in its own terns can be regarded as a simple solution, although it certainly offers a much wider choice. DRA (1992) suggests that over time, there will be a switchover to electricity as households acquire electrical equipment and extend electricity usage in the household. However, entry costs (especially in the DFR) and usage costs may limit access for the lowest income households, many of which will be woman headed. Further, gendered relations and practices may continue to shape energy usage so that women do not experience the benefits assumed to be associated with electrification. Policies like electrification do not challenge the gender division of labour in terms of which domestic work is women's work. At best, they may meet women s "practical gender needs , but as we have shown, the position of women in the economy and in the household limits the extent to which even these needs are met.

4.4 Transport:

Studies on infrastructure in both urban and rural areas point to spatial access as a critical issue. It is common cause that the extensive homework distances prevalent in South African cities contribute to economic inefficiency through their impact on productivity, and the high expenditure on transport by both the state and users. Studies of women and transport in SA cities suggest that these difficulties may be exacerbated for women, and particularly for certain low income groups. Women have to negotiate between the demands of work, caring for children and domestic work - and the different spatial locations of these activities. Cook (1987) shows the time demands created by the need to negotiate between these various activities, and the stress it places on women. Many women in her study of Khayelitsha in Cape Town had very little leisure time and were forced to sacrifice sleep to accomplish the various tasks and to commute to work each day.

The type of work that women are engaged in in cities may present particular difficulties from a transport perspective. There is a paucity of studies which document the transport needs of particular groups of women. Nonetheless, it is likely that the kind of work significant groups of women are engaged in and their location exacerbate their transport problems. For example, a large proportion of African women are domestic workers who will face transport difficulties due to the poor public transport networks in wealthier areas. The hours and location of nurses, who are another major group of African women workers, are also likely to be off the mainstream. Part-time and shift workers many of whom are probably women - may also experience transport difficulties as they have to move outside of peak times.

Naidoo (1993) shows that the costs of transport, given the very low incomes of women informal traders, and their long working hours, has meant that a number of women in the Warwick Triangle area have begun to sleep on the streets near to where they work, although they may have a home in the city. A number of groups of mainly women, generally involved in informal work, have also begun to sleep on the streets of the Durban central area. There is little research on this issue, although it seems likely that the transport cost-distance factor, and the nature of the work, will be significant factors in accounting for this trend.

In rural areas, spatial marginality limits women s mobility as well as their access to facilities and economic opportunities, and as a result, it drives up the cost of living. The poor condition of many local roads and their lack of maintenance (May, 1993b) has meant that transport costs are pushed up, and that taxis avoid going into certain areas, effectively isolating these areas.

4 5 Health:

Access to health services, and conditions underpinning health, are of importance to women not only from a personal perspective, but because the work involved in caring for the sick at home generally falls on women's shoulders. Many women are employed in jobs where there are no health/medical aid benefits, and which are not covered by legislation regarding health conditions. They are more likely to have to rely on public services than are men.

Patterns of public health provision generally follow patterns notable in other sectors. Urban areas in general have much better access to mobile and permanent health services than rural areas. Services in KwaZulu are generally better than in Natal due to years of state policy which denied Africans permanent residence in Natal. Hambridge and Krige (1991) show that while virtually all of the urban African population has access to facilities, some 30% of the rural African population in Natal and 22% in KwaZulu are more than 5 km from mobile services. Access to permanent facilities is worse, and "much of rural Natal is a veritable desert as far as permanent health facilities are concerned" (Hambridge and Krige, 1991, 69).

While access is better in urban areas and in KwaZulu, it is still not adequate. Health services in many areas are strained by the numbers seeking attention relative to the facilities and personnel available. In rural areas, the availability of only mobile services in a number of areas, and the distances to more permanent facilities, is often a source of complaint. A particular concern raised by women in this regard is the lack of maternity facilities and their unavailability on a 24 hour basis.

4.6 Educare:

A key 'resource affecting women s participation in the paid economy is access to adequate child care. Women are generally seen as responsible for caring for children, and it is impossible for them to engage in paid work outside of the home if alternative arrangements for caring for children cannot be made. In some countries the state has played a significant role in providing childcare facilities, or in monitoring and supporting them. In South Africa, childcare has tended to be marginalised in state expenditure policy. For example, Liddell (1990) shows that less than 1% of DET funding goes to preschool education. There are limited monies available for either the development of preschool facilities or for operational costs. Consequently, the cost of care for preschool children falls on families. As usual, there are racial differences in state expenditure on preschool education, with double the number of white children in state registered preschool centres as black children.

Within Natal/KwaZulu, less than 5% of children between 0 and 6 years old are at an educare facility (Biersteker et al, 1993). Provision for under 3 s is particularly poor (Desmond, 1993). Most African women rely on relatives, neighbours or unregistered childminders. Studies suggest, however, that these solutions are seen as unsatisfactory, largely because individuals are less reliable than are institutions. Institutions are also seen as preferable from an educational and safety perspective (Liddell and Kemp, 1991). Furthermore, the cost of sending children to an institution may be lower than paying a childminder - if payment is involved - because institutions effectively can experience "economies of scale in childcare. It is interesting to note that educare training institutions report a growth of local creches, and growing demand for them as one of the first development priorities in some communities primarily so that women can go out to work (Desmond, 1993). This observation certainly matches evidence of the increasing participation of women in the labour force.

Educare may also be an important, but undocumented, source of largely women s employment. Moller (1988) reports that it accounted for as much as a third of informal employment in her survey of KwaMashu. Care of children might be rewarded in indirect ways! such as through providing meals and other informal exchanges, rather than through direct payment. Unfortunately, there is no systematic survey of the numbers of women involved in this kind of employment or of their remuneration, although these figures may well remain unrecorded in many surveys. There is also no systematic information on unregistered educare centres. Education NGOs, however, note that wages for teachers are extremely low and not very different from the returns to informal work - around R100 to R300 pm (Goldberg, 1993; Desmond, 1993) - because of the amount that households in communities can pay (R15-R35 pm). Conditions in rural areas are worse as incomes are far lower, but the costs of operation are as high as in urban areas.

4.7 Concluding comments:

Access to the range of physical and social infrastructure discussed above has important implications as we have shown for the extent of unpaid work, and for women's ability to participate in paid work. Conditions around access to resources are probably worse in the region compared to many other parts of the country. There are proportionately more people in informal settlements, and infrastructural deficits are worse than elsewhere. Consequently the conditions around women s unpaid work are probably worse. While a feminisation of the labour force is occurring, there is no evidence of an equivalent "masculinisation" of domestic work or of a changing sexual division of labour within the home. Rather, other women who are not employed or working elsewhere, are picking up domestic work in the home. Where such supports are unavailable, the pressures on women must be enormous.

CONCLUSION

One of the most important trends which this study of women and gender in region E has highlighted is the growing feminisation of the labour force. This process appears to be largely the result of the growth of employment in sectors in which women are dominant, although some gender shifts in sectors are notable as well. The growing presence of women in census figures is also the result of stagnation, and in some cases, the decline in men s employment.

The feminisation of the labour force, however, is not necessarily a panacea for women, nor is it an unproblematic process. As we have shown, women are still a smaller part of the labour force than men. Despite migration to cities, about half of all women in the region are still concentrated in rural areas where opportunities are extremely limited. Furthermore, women are crowded into certain parts of the labour market, and predominate in low-wage, less secure jobs and in the informal sector. In consequence, women's incomes are substantially lower than those of men, and women headed households form a disproportionately large part of the poverty groupings.

The feminisation of the labour force is also occurring under difficult conditions for women. As we have shown, poor social and physical infrastructure greatly increases the burden of unpaid domestic labour, which is generally the preserve of women. Conditions in Region E are particularly poor. A large proportion of households in urban areas live in poorly serviced informal settlements. Where working women are unable to rely on the support of others to do domestic work, the stress and pressure on them must be enormous.

Restructuring of the economy therefore needs to occur in a way which is sensitive to the connection between individuals and household survival, and the kinds of employment which will be made available. In this regard, it wil not be possible to address fundamental issues of inequality or equity without seriously considering the growing feminisation of poverty and the vast numbers of people and households who are dependent on women s earnings.

While we have focused largely on the feminisation of the labour force from the perspective of women, the impact of stagnation or decline in men s employment is also significant. From men s perspective, the rising unemployment crisis may be greater than the average figures suggest. It may well be the case that these shifts are contributing to the spiralling violence in the region. Campbell (1989), for example, shows how men s unemployment, and their frustration translates into violence within the family, and often into violence against women.

Our discussion of women s increasing employment, and the conditions under which it is occurring, raises questions about the gender implications of policy thinking within the Main Summary Report for this study (Harrison, 1993). It is not possible to undertake an exhaustive commentary on the report. We will, however, point to some areas in which a gendered analysis raises questions about policy directions discussed in the report.

A great deal of attention is placed in the report on restructuring the region to make it more competitive internationally. There is a focus on enabling the region to adopt the new flexible technologies which have become prevalent internationally. It is impossible to say what the impact for men and women of this type of restructuring will be, nor is it possible simply to transpose conclusions from the international context. Nevertheless, it is worth noting that internationally, the trend in flexible production processes has generally been to divide the work force into a small core of skilled and generally male labour, and a larger peripheral flexible and mainly female labour force. If South Africa does follow the international trend, women s employment will not necessarily shift from the classic focus on largely low wage, less secure work.

Regional restructuring along the lines suggested in the Main Report may well lead to a restructuring of the space economy of the region as different places become desired locations and as different kinds of labour forces are sought. Some of the present decentralisation points may become vulnerable in this process. As Harrison (1993) notes, the changed Regional Industrial Development Programme has already had a differential effect over space, and some places previously supported are now vulnerable. As we have shown, the contribution of some of these areas to women s employment is not insignificant and they cannot simply be disposed of. Development strategies for these areas may be as necessary as they are for coal mining areas. In developing strategies, it will be important to take into account what kind of jobs are created, and who is likely to occupy them.

Small business is posited as an area where both the aims of growth and redistribution can be achieved. As we have shown, while women appear to be employed to a large extent by small business, very few are owners. The benefits of a small business development strategy are likely to be skewed in favour of men unless specific efforts are made to ensure that women benefit. Research needs to be undertaken to establish where women are located within small business and the barriers and constraints on their participation as owners.

Similarly, smallholder agriculture is seen as a potential way of generating both growth and redistribution. But a smallholder policy which does not consider gender issues could reinforce present inequities and thereby undermine its goals. This study has shown that the way rights to land are defined in homeland areas pose particular problems from the perspective of women engaged in smallholder production. Where women are married, men own the land and therefore have control over the proceeds of the crop, even though women are the producers. This system affects women s access to the income they have earned, and can impede production. If women are to benefit from a smallholder development strategy, attention will have to be given to issues of rights to the land and to the income from a small holder strategy.

The development of infrastructure is an obvious strategy for the region as Harrison (1993) notes. There is the potential to stimulate the depressed construction industry, to generate employment through public works schemes, and to meet basic needs. The strategy is positive from a gender perspective as it has the potential to reduce substantially women s unpaid domestic labour. In practice, however, the ability of an infrastructure development programme to benefit women will depend very much on how it is implemented.

Affordability is a mayor concern in infrastructure development programmes, but gender relations within communities and households can also have very significant effects on women s access to the benefits of programmes. As we have shown, electrification has been of far less assistance to women than might have been assumed. Similarly legal definitions regarding access to subsidised sites may disempower married women and prevent their access to these sites on divorce.

Since the construction industry is largely the preserve of men, an infrastructure development programme may marginalise women in employment creation if no steps are taken to ensure that they can get jobs in these schemes.

Finally, little mention is made of the largest formal employment sector for women - the community services sector. The community services sector is not only important from an employment perspective, but also because it encompasses a range of social services which affect women s work. Policies to privatise social services, especially health and welfare services, may increase the work of women, who traditionally are responsible for care of the sick, the disabled and so on (Lund, 1993). Conversely, improvements in social services, and in welfare payments may have very significant positive effects for women and women headed households. Given the importance of domestic work in women s employment, more research needs to be undertaken to understand it as an economic activity in the region, and what strategies might be adopted.

In considering gender in relation to regional economic development strategies, a broader approach is also necessary Strategies will have to be developed to challenge the sexual division of labour which limits women s access to employment and confines them to relatively lower paid work. These might include affirmative action policies, skills training programmes to ensure that women move up the skills and wages hierarchy, as well as others. Education programmes may also be necessary to overcome the unwritten barriers and taboos regarding women s work and the kind of positions they can occupy. Without such programmes in place, it is likely that very few women will be able to take advantage of changing opportunities.

An integral aspect of the challenge to gendered roles, and assumptions around gendered abilities, concerns the question of unpaid labour, and particularly that labour crucial to the process of social reproduction. Women s labour in the home, and in areas of reproduction, is invisible as a form of work. One way in which the nature and extent of unpaid labour could be recorded, and incorporated into economic policy, is through quantifying this labour. Quantifying domestic labour will not provide the whole solution to removing the gender bias in all aspects of economic activity. But it will form an important part of that process which seeks to fundamentally review current economic thinking and policy making. The measurement of unpaid reproductive labour will help challenge the existing bias of economic statistics, prevailing conceptions of economic activity, and the current focus of economic policy.

REFERENCES

Annecke W A (1992), We Are So Poor. Unpublished MA Thesis, University of Natal, Durban.

Ardington E M (1984) "Decentralised industry, poverty and development in rural KwaZulu". Carnegie Conference Paper 248.

Ardington E M (1988) "Nkandla Revisited: A Longitudinal Study of the Strategies adopted to Alleviate Poverty in a Rural Community". Working Paper no 16, CSDS, University of Natal, Durban.

Ardington E M (1991) "Education for All. Spatial Dimension". Indicator. vol 8, no 2.

Baldogh C (1993) "Economic Development Study for the Zululand Subregion. RDAC report on region E.

Becker G (1980) "A Theory of the Allocation of Time". Amsden A (ed) The Economics of Women and Work. England: Penguin Books.

Biersteker L et al (1993) "Early Childhood Educare in South Africa A Situational Analysis", ms.

Bonnin D, Friedman M, Hassim S. Todes A, Vaughan A (1991) "State, Gender and Restructuring in South Africa in the 1980s". Paper presented at the "Conference on Women and Gender in Southern Africa", University of Natal, Durban, 1991.

Bozzoli B (1991) "Life Strategies, household resilience and the meaning of informal work: some women's stories". Paper presented at the "Conference on Women and Gender in Southern Africa", University of Natal, Durban, 1991.

Brooks S and Friedman M (1991) "Community Gardens in KwaZulu: A Pilot Study". Institute of Natural Resources. Investigational Report no 68.

Budlender D (1991) "Women and the Economy". Paper presented at the "Conference on Women and Gender in Southern Africa", University of Natal, Durban, 1991.

Budlender D, Hendrie D, Homer D and Young G (1984) "Industrial Council Wages Rates and Poverty". Carnegie Conference Paper, no 107.

Buijs G (1985) "Women on their own: Migrants from the Transkei employed as agricultural laboureres in Natal. Paper presented to the Annual Conference of South African Anthropologists, Durban.

Campbell C M (1989) Township Families and Social Change: A Pilot Study. Unpublished MA Thesis, University of Natal, Durban.

Carr C (1991) Personal communication.

Cellier G (1993) Personal Communication.

Central Statistical Services (1980) Population Census (after adjustment for undercount) 03-80-05; 03-80-08; 03-80-10.

Central Statistical Services (1992) Population Census (after adjustment for undercount), nos: 03-01-01; 03-0103; 03-01-09; 03-0110

Cook G (1987) "Town Budgets of Working Women in a Disadvantaged Society". Blue Stocking.

CORD (1990) "Overcoming Apartheid's Land Legacy in Maputaland (Northern Natal". Working Paper no 3.

Cross C, Nzama M and Dlamini O (1983) "When the Rains Came: Women of Thousands Hills in the 1987 Flood Disaster". Africanus, 18: no 1 and 2.

Cross C, Bekker B and Clark C (1992) "Moving on: Migration Streams in and out of Inanda". Centre for Social and Development Studies, University of Natal, Durban.

Data Research Africa (1992) "An In-depth Evaluation of Fuel Use by Rural Women. The Ingawavuma District - KwaZulu". Durban.

Data Research Africa (1993) "Eskom Post Electrified Area Research". vol 1, Durban.

Desmond S (1993) TREE, personal communication.

Elson D (1991) "Gender Analysis and Economics in the Context of Africa". Manchester Discussion Papers in Development Studies.

Friedman M (1991) "Commercial Timber Production and Basic Needs: Economic and Organisational Potential of Small Growers in KwaZulu lessons from an Integrated Rural Development (IRD) Project. Institute of Natural Resources Occasional Paper.

Friedman M and Hambridge M (1992) "The Informal Sector: Gender and Development". Preston-Whyte E and Rogerson C (eds) South Africa's Informal Economy. Oxford University Press, Oxford.

Gandar M V and Bromberger N (1984) "Economic and Demographic Functioning of Rural Households: Mahlabatini District, KwaZulu. Social Dynamics, vol 10, no 2.

Ginwala F. Mackintosh M and Massey D, (1991) "Gender and Economic Policy in a Democratic South Africa". Development Policy and Practice Research Group, Working Paper no 21.

Goldberg S (1993) Urban Foundation. Personal Communication.

Gwagwa N (1993) "Low income Household Resource Allocation. The Case of New Town, Durban. Unpublished Preliminary Results.

Hambridge M and Krige D (1991) "Health Care. In Search of a Cure". Indicator, vol 8, no 2.

Harris N (1991) "Productivity and poverty in the cities of developing countries". Paper presented at the Conference on the Urban Agenda for the Nineties and Experience of National Urban Policy.

Harrison P (1993) "Economic Development Strategies for Region E". RDAC summary of reports on region E.

Hellman E (1974) "African townswomen in the process of change". South African International 5.

Hofmeyer J (1993) "Labour in Region E" RDAC report on region E.

Hulbert S (1991) Personal communication.

Jacobs M (1990) "A Statistical Overview of Education in KwaZulu Natal 1990". Edusource, Johannesburg.

Jaffee G (1988) "Commuter Labour: Changing Women's Lives, Changing Households", Agenda 3.

Jenson J. Hagen E (1988) "The Feminisation of the Labour Force. Paradoxes and Promises. Jenson J. Hagen E and Reddy C (eds) Feminisation of the Labour Force. Oxford: Polity Press.

Kridge D (1989) "The basic needs approach to development: The question of education for black people in Natal". Development Southern Africa, vol 6 no 2.

Krige D (1990) "The basic needs approach to development: The question of health care for black people in Natal". Development Southern Africa. vol 7, no 1.

KwaZulu Finance Corporation, nd. Special Report, Investments Department.

Liddle C and Kemp J (1991) "Current Issues for Day Care in South Africa". Early Education Research Unit, HSRC.

Lines L (1990) "Informal Networks for Pre-School Children in a Black Community in South Africa". Community Development Journal. vol 25, no 4.

Lyne (1991) Personal Communication.

Maconachie M (1989) "Looking for Patterns of Women s Employment and Educational Achievements in the 1985 Census". Agenda 5.

Maconachie M (1993) "Patterns of Women's Employment in the 1991 Census". Agenda 18.

Marcus T (1985) "Women Farm Workers in South African Commercial Agriculture". Unpublished paper, Holland.

May J (1987) "The Composition of the Black Labour Force: Spatial Demographic and Gender Differentiation in KwaZulu". Paper presented to 18th Conference of the Association of Sociology in Southern Africa. Western Cape.

May J (1993a) "Demography and Social Concerns in Region E". Data Research Africa.

May J (1993b) "Socio-Economic Processes in the Rural Areas of Region E". RDAC report on region E.

May J (1993c) "Commerce as an Economic Sector Within Region E. The Informal Sector and Micro-Enterprises". RDAC report on region E

May J and Rankin S (1990) "The Spatial and Gender Differentiation of KwaZulu's Black Labour Force". Agenda 6.

McCarthy J (1993) Personal Communication. Referring to his unpublished work on industry in the DFR.

McIntosh A (1992) "Making the Informal Sector Pay. Rural Enterprise in KwaZulu Preston-Whyte E and Rogerson C (eds) (1992) South Africa's Informal Economy. Oxford University Press, Oxford.

McIntosh A and Friedman M (1989) "Women' Producer Groups in Rural KwaZulu". Development Southern Africa) vol 6, no 4.

Moller V (1988) "Community Reactions to the Introduction of an Educare Programme at a Home for Black Aged in Kwamashu". Centre for Applied Social Sciences, University of Natal, Durban.

Naidoo K (1993) Work and Life of Women in the Informal Sector: A case study of the Warwick Avenue Triangle. MA Thesis, Geography, University of Natal, Durban.

Parnell S (1991) "Race, class, gender and housing subsidies". Urban Forum. vol 2, no 1.

Pentz D C (1993) Informal Street Trading: a case study of Warwick Avenue. MA Thesis, Town Planning, University of Natal, Durban.

Pillay P (1985) "Women in Employment in South Africa: Some Important Issues and Trends". Social Dynamics, vol 11, no 2.

Posel D (1993a) "Some Notes on Quantifying Domestic Labour". Paper presented at the Unpaid Labour Workshop. Commission on the Emancipation of Women and the ANC Research Department, Johannesburg.

Posel D (1993b) "The Debate Over Prostitution. The Effects of Criminalising Sex Work and the Search for Alternatives". Occasional Paper, Economic Research Unit, University of Natal, Durban

Preston-Whyte E (1992) "'Invisible Workers': Domestic Service and the Informal Economy".

Preston-Whyte E and Nene S (1984) "Where the Informal Sector is not the answer: Women and poverty in rural KwaZulu". Carnegie Conference Paper, no 235, Cape Town.

Preston-Whyte E and Nene S (1982) "Black Women and the Rural Informal Sector". Preston-Whyte E and Rogerson C (eds) (1992) South Africa's Informal Economy. Oxford University Press, Oxford.-

Preston-Whyte E and Rogerson C (1992) South Africa's Informal Economy. Oxford University Press, Oxford.

Preston-Whyte E and Zondi M (1989) "To Control their Own Reproduction: The Agenda of Black Teenage Mothers in Durban". Agenda 4.

Pudifin C and Ward S (1986) Gender and Decentralisation in Isithebe. Unpublished M.A. Thesis, University of Natal, Durban.

Simkins C, (1986) "Household Composition and Structures in South Africa". Burman S and Reynolds P (eds) Growing Up in a Divided Society. Johannesburg: Ravan Press.

Smith G (1993) "An Analysis of the Provision of Physical Infrastructure and Delivery of Services in Region E". RDAC report on region E.

Sneddon H (1990) The Clothing Sellers of Warwick Avenue: Getting to Know Women in the Informal Sector. BA Hons Thesis, Department of Sociology, University of Natal, Durban.

Stavrou S (1992) "Electrifying the Nation". Indicator SA, vol 3, no

Taylor D (1985) "Women: An Analysis". Brekke T et al, Women. A World Report. London: Methuen.

Todes A and Smit D (1993) "Urbanisation Study". RDAC report on region E.

Todes A and Walker N (1992) "Women and Housing Policy in South Africa: A Discussion of Durban Case Studies". Paper presented to the Conference on Women and Shelter: First and Third World Perspectives, University of Michigan, Ann Arbour.

van der Vliet (1982) Black Marriage: Expectations and Aspirations in an Urban Environment. Unpublished M.A. Thesis, University of Witwatersrand, Johannesburg.

van Zyl J. von Bach H S and Kirsten J (1993) "The Agricultural Sector in Region E". RDAC report on region E.

Vaughan A (1992) "Options for Rural Restructuring". Schrire R (ed) Wealth of Poverty? Critical Choices for South Africa. Oxford University Press, Oxford.

Vaughan A (1993a) Personal Communication.

Vaughan A (1993) "Farmer Support Services Supplied by NGOs and Commercial Enterprises: The Sugar Industry". Paper prepared for the World Bank.

White C (1991) "'Close to home' in Johannesburg: sexism in township households". Paper presented at the Women and Gender Conference in Southern African, University of Natal, Durban.