Women in the urban labor markets of Africa : the case of Tanzania

Women in the Urban Labor Markets of Africa: The Case of Tanzania SWP380 World Bank Staff Working Paper No. 380 I April 1980 lb~~   The views and interpretations in this document are those of the author and should not be attributed to the World Bank, to its affiliated organizations, or to any individual acting in their behalf. WORLD BANK Staff Working Paper No. 380 April 1980 WOMEN IN THE URBAN LABOR MARKETS OF AFRICA: THE CASE OF TANZANIA The low status of women in the modern sector in Africa has been commented on by many authors but very few studies have sought to explain the reasons for the disparity in employment opportunities in the modern sector between men and women in Africa. This study seeks to fill the gap in the literature by using the results of the 1971 Tanzanian National Urban Mobility, Employment and Income Survey (NUMEIST) to focus specifically on urban women and their labor market performance. The study examines the structure and characteristics of the female urban labor force, the determinants of female employment and labor earnings and attempts to ascertain whether sex differ- ences in labor market performance can be attributed to sex discrimination. Apart from the constraints arising out of the general economic situation in the country, factors that emerge as important determinants of female employment status are: marital status, nature of female dependence on the husband and educational attainment. The analysis of differentials in labor earnings indicatesthat, while African urban women may not suffer appreciably from sex discrimination, either in terms of wage level or as a result of occupational segregation, the urban women's low economic status in the labor force is primarily a result of cumulative discrimination over time in both the provision of and the demand for educational services. This has had the effect of limiting the supply of women in the urban labor force and making the female labor supply curve less elastic than that of men. The paper discusses some policy options that governments may wish to adopt to remedy the situation. Prepared by: Nwanganga Shields Population and Human Resources Division Development Economics Department INTERNATIO;O?AL 0R=ETARY U FET P 2;) ; ___ .a ;r.TL Orl irlr  Copyright C 1980 International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. First printing April 1980 Second printing May 1982 Third printing June 1984 All rights reserved Manufactured in the United States of America  Acknowledgment I wish to thank Richard Sabot for making available to me the Tanzanian NUMEIST Tape; Oey Meesook, Gloria Scott, Anna Sant'Anna, Jack Maas and Richard Sabot for useful comments on an earlier draft; Mike Hartley for advice on methodology, and Judy Bobb for computer help. Any remaining faults are entirely mine.   Table of Contents Page No. Chapter 1 Introduction and Summary .............................. 1.1 Objective ......... o............... o............. oo..-.. 1.2 Data Source .....o..........o.... ...................... 1.3 Chapter Summary ...... .-...... .... ........ ......... 3 Chapter 2 Urbanization of Women in Africa -....... ....... 8 2.1 Urban Growth in Africa ...............000.............. 8 2.2 Urbanization of Women ..................................... 17 2.3 Summary oo ................... o................... o ... o .... oo ... o.o. 27 Chapter 3 Characteristics and Structure of the Female Urban Labor Force ..o..ooo ..... ......... o. . ..... 28 3.1 Introduction o .... oo.. ..o...o-oo.. 28 3.2 Socioeconomic Characteristics of the Urban Females .... 29 3.2.1 Demographic Characteristics .9.................. 29 3.2.2 Education 34o...................o.......... oooo.. 34 3.2.3 Employment Status of Women in the Urban Labor Market 36 3.3 Occupational Distribution of Women ................ . 43 3.4 Summary o....o." - -oo..o -oo- ... 54 Chapter 4 Determinants of Female Employment ... ........ 55 4.1 Introduction . .. . .... . .... . . ... 55 4.2 Determinants of Female Employment 57 4.2.1 Life Cycle Commitment 57 4.2.2 Education and Work Experience 68 4.3 The Model and Empirical Estimation of the Labor Supply -...o.. . ... o.o...o...o... 76 4.3.1 Hypothesized Relationships .... o.......0000.000".0 80 4.3.2 Description of the Results ...o...o..oooo....ooo...ooo. 84 4.4 . Summary voooooooooo*o*oo.**oo*o 8t7  Table of Contents (Continued) rate No. Chapter 5 Determinants of Earnings ........... ................. 88 5.1 Introduction ...................... 88 5.2 Education, Experience and Earnings ................... 89 5.3 The Estimation of the Determinants of Earnings ........ 99 5.3.1 Empirical Model ...... *..i..... oee- 99 5.3.2 Results of Analysis ...o.-.. ......-... . . . .... ..... .. . 102 5.4 Possible Explanation of Sex Differences in Earnings and Labor Market Involvement in Africa ..... 0........... 104 5.4.1 Review of Existing Literature ......................... 104 5.4.2 An Alternative Explanation . ............................. 110 5.4.3 Relevance to Africa ...... ...- . ...... ....-........ 111 5.5 Summary ..oo..o ..... 00..0.... ..... .O...00.......... 0 124 Chapter 6 Policy Implications ......e.......................... 125 Appendices A.1 Marital and Household Status of the Urban Population by 5-year Age Cohort A.2 Female Population by Marital and Household Status: Age Structure A.3 Level of Education by Sex and Age Group A.4 Estimated Total Fertility and Birth Rates for Tanzania Based on the Female Crude Birth Rates, 1967 A.5 Female Population: By Age and Number of Children  CHAPTER 1 INTRODUCTION and SUMMARY 1.1 Objective Despite considerable literature on urbanization and the problems associated with surplus urban labor in Africa, few attempts have been made to study the process of urbanization of women and the performance of women in the urban labor market. It may be argued that urban problems of under util- ization and their principal forms, under-employment and unemployment, are shared by men and women and that policy prescriptions arising out of research on unemployment as it affects men are also applicable to women, but many of the problems women encounter in the urban economy and the labor market are sex specific. Some of these problems are associated with their life cycle commitments, while others arise because of the low status of women in society. A man at the end of his educational training is expected to enter the labor force and to remain there, assuming that there are jobs available, until retirement; a woman's entry into the labor force is conditioned by her plans relating to marriage and childbearing, as well as by the social expectations of what role she should assume in society. In Africa, although society expects her to enter the labor market, she is still expected to fulfill her role relating to procreation and family responsibilities. While labor market conditions in the rural sector may make it possible for a woman to pursue dual careers (and most African women do so), conditions in the urban areas are less conducive than in rural areas to the combination of home and market activities. Furthermore, in the urban environment, where the necessity for a woman to work is even more pressing for some families owing to the husband's low income and increased family wants, women have more severe problems in adjusting to unfamiliar life styles.  Urbanization is a fairly recent phenomenon in Africa and most urban dwellers are first generation or recent immigrants. The inability to draw on the support of the wider extended family so readily available in the rural areas affects women more intensely than men. The low status of women in the modern sector in Africa has been commented on, but very few studies have sought to explain the reasons for the disparity in employment opportunities in the modern sector between men and women. This study seeks to fill the gap in the literature by focusing specifically on urban African women and their labor market performance. A subsidiary but important reason for a study focusing specifically on urban women is the importance of female employment as a means of providing women with an alternative source of satisfaction to that derived from having children. In high-fertility African countries, where female status derives almost entirely from the motherhood role, many women feel compelled to have as many children as possible. Research results indicate that modern sector employment can offer women viable partial alternatives to marriage and motherhood. Thus, studies which seek to shed light on the employment status of women in urban areas of Africa should be useful to policy makers. The objectives of this study therefore are to ascertain the char- acteristics of the women in the urban labor market and to compare the labor market performance of men and women in the urban economy. Some of the ques- tions we hope to address are: Are there sex differentials in employment status and earnings? If differentials exist, to what factors can they be attributed? Are they explained by differences in education and job train- ing, or are they even a result of discrimination? And finally, what steps might be taken to close the gap between the sexes and help further integrate women into the urban economy?  -3- 1.2 Data Source As a principal source of information, Tanzania has been selected for this study because there are fairly good household data on employment and income in the urban areas. The 1971 National Urban, Mobility, Employ- ment and Income Survey (NUMEIST) provides a rich source of data suitable for an analysis of this type. The survey covered 5,000 persons from seven urban centers in Tanzania. These are: Dar es Salaam, Arusha, Mwanza, Tanga, Tabora, Dodoma and Mbeya. The sample selection and enumeration procedures are described by Sabot. 1/ The other reason for concentrating on Tanzania is that in the Arusha Declaration, 1967, the Tanzania Government committed itself to a policy of sexual equality and has since taken considerable steps to ensure that both urban and rural women benefit from this policy. This study is not restricted to Tanzania, but has tried where possible to include information from other African countries. While it can be said that each country's position is fairly unique, we feel confident that the analysis and the conclusions of this study are applicable to the majority of sub-Saharan African countries. 1.3 Chapter Summary Previous studies have shown that rural-urban migration in Africa has been predominantly male and until fairly recently there has been very little female migration into urban areas of sub-Saharan Africa. In the past, given the short-term nature of male urban migration, it was most unusual for wives to accompany husbands to town. However, in the post-second world war period, there was a sharp increase in the size of the female 1/ R. Sabot: Economic Development and Urban Migration, A Study of Tanzania. Clarendon Press 1978.  - 4 - urban population as indicated by the sex ratios of major African urban areas. Since the increase cannot be explained by the natural rate of increase alone, it indicates a shift in the pattern of female migration. Chapter 2 discusses the reasons for this phenomenon. There was a shift in male migra- tion patterns from short-term to long-term during this period. Consequently, the move to the city involved whole families instead of just the male member as in the past. During the post-war period, the rate of male migration into urban areas increased sharply, owing to increased economic opportunities in the modern sector. Since the rate of female migration derives to a large extent from male migration, it was inevitable that this would have an effect on the rate of entry of female urban migrants. Although the proportion of female migrants who came as economic dependents remained high, the rate of migration of unattached females grew at a much higher rate than that of married females, indicating that women migrants were also responding to the same economic motives for migrating as the men. A factor contributing to this was education. During the 1960s, there was a net increase in the number of women in educational institutions in Africa. Given the educational selectivity of the migration process in Africa, it was not surprising that the process should also affect women. To the educated woman, the urban economy offered the probability of employment towards which school curricula have been directed, as well as the oppor- tunity for upward social mobility. Given the high female labor-force participation in rural Africa and the numerical importance of women in the urban population (50% of urban work age population are women), Chapter 3 examines the structure and char- acteristics of the female urban labor force. It begins with an analysis  of its socio-economic characteristics. The important point to emerge is that although women make up approximately 50% of the urban work age popula- tion, there are very few women in the formal sector of the urban economy, where educational attainment is a pre-requisite. Using the broadest defini- tion of employment, we find that in Tanzania, approximately 21% of the urban women are employed compared with 80% of the men. The majority of women urban labor force participants are in job situations not very different from that of rural women. An examination of the occupational distribution of women indicates the beginning of a pattern of concentration of women in few occupa- tions similar to the pattern in developed countries, although the pattern has yet to become as pronounced as in Kenya. Chapter 4 examines the determinants of female employment in Africa. Apart from constraints arising out of the general economic situation in the country, factors that emerge as important determinants are marital status and the nature of female dependence on the husband, the presence of children and educational attainment. The chapter hypothesizes that education by itself cannot ensure access to employment, but that the chances of being employed are greatly improved by the nature and length of a woman's on-the-job experience, especially in a tight labor market situation. Finally, a labor supply func- tion using household survey data from urban Tanzania was estimated. The results of empirical analysis confirm the importance of wage, education and household status as determinants of labor force participation. Moreover, education is an important factor in determining whether or not a woman finds employment in the formal sector. Given the wide differentials in earnings and labor force status of men and women in urban Africa, Chapter 5 examines the determinants of  - 6 - labor earnings and attempts to ascertain whether sex differences in labor market performance can be attributed to sex discrimination. Important reasons for differential earnings are women's low level of human capital and skills which preclude them from high-level jobs with high remuneration. The preponderance of women in the informal sector of the urban labor market can be explained by the fact that education is not a prerequisite for entry into employment in this sector. However, since high remuneration in this sector is dependent on capital outlay and many women lack this, their prospects of earning high incomes are severely limited. The empirical estimation of the earnings function confirms that women do not suffer from sex discrimination, but that labor income is highly correlated with educational attainment. This leads to a discussion of the reasons for African women's comparatively low educational attainment vis-a-vis that of African men. The conclusion from the evidence presented is that, while there may not be sex discrimination in urban Africa either in terms of wage level or occupational segregation, women's low status in the labor force is a result of cumulative discrimination over time in both the provision of and the demand for educational services. This has had the effect of limiting the supply of women in the labor force and making the female labor supply curve less elastic than that of men. Supply differences between men and women are determined by family decisions and have been reinforced by Government action. Expansion of female skilled labor has been prevented by discrimina- tory practices of fathers who do not give their daughters access to education similar to that of their sons and by the governments who do not provide sufficient training facilities to women. Chapter 6 contains the policy implications.  - 7 - Policy suggestions that have emerged from this study are as follows: (1) Girls should be provided with more access to education by making all schools coeducational and ensuring that the school intake reflects the ratio of the sexes in the population and also that girls are not automatically put into an inferior educational stream which would not give them a chance of competing in high status and high income jobs. (2) Major efforts should be made to expand and modify the adult education system, so that the current illiterate, underemployed and poor urban women may have an oppor- tunity of upgrading their skills and ability and thereby increasing their income. (3) The educational system should also be restructured so that provision is made for the school dropouts and older people to re-enter the system and obtain education which would improve their earning potential.  - 8 - CHAPTER 2 URBANIZATION OF WOMEN IN AFRICA 2.1 Urban Growth in Africa Africa is undergoing tremendous structural change and one of the most striking characteristics of this change is urbanization. The process of urbanization is all the more interesting because when compared with de- veloped countries, and even the less developed countries of Asia and Latin America, Subsaharan Africa has relatively low urban density. Estimates made by Davis 1/ in 1968 indicated that approximately 10% of the population of Subsaharan Africa lived in urban centers of 100,000 or more persons. One important aspect of urbanization in Africa is that it is fairly recent. Most of the increase has occurred during the second half of this century. Avail- able statistics indicate that in each of the 1950-60 and 1960-70 decades, the increase in urban population in Africa was about 60%. Longitudinal data on individual towns are not easily available but such information as exists shows that most cities experienced an annual growth rate of about 5% during the 1960-1970 decade. Of the cities listed in Table 2.1, Koidu town in Sierra Leone, Arusha in Tanzania and Nairobi in Kenya had an estimated annual growth rate of about 10%. Intercensal data from Tanzania and Kenya on population change of major towns are indicative of the tremendous growth of most towns of Sub- saharan Africa during the second half of this century. Table 2.1(a) gives 1/ K. Davis: World Urbanization, Population Monograph Series No. 9, Uni- versity of California at Berkeley 1972.  Table 2.1: GROWTH RATE OF MAJOR TOWNS OF SELECTED AFRICAN COUNTRIES CIRC. 1950 - 1970 Country Annual Percentage Growth Rate and City 1950-60 1960-70 Ghana Accra * 8.4 6.8 Kumasi 5.8 Nigeria Ibadan 2.2 3.0 Lagos * 4.6 6.7 Kano 5.0 5.0 Abeokuta 3.8 4.8 Ilesha 8.6 Ivory Coast Abidjan * 7.6 Kenya Nairobi * 9.7 (1962-69) Mombasa 4.7 Kisumu 4.7 Tanzania Dar es Salaam * 7.5 (1948-67) Mwanza 6.1 Mbeya 6.7 Tanga 5.9 Arusha 10.0 Sierra Leone Freetown * 6.5 (1963-74) Koidu Town 15.6 Source: Ghana, Nigeria, Ivory Coast: K. Davis: World Urbanization 1950-70. Vol. I Population Monograph Series No. 4 University of California at Berkeley, 1971. Kenya - calculated from 1962 and 1969 census. Sierra Leone - from C.S. Okoye: Migration in Sierra Leone IBRD mimeo 1975 Tanzania - as in 2.la * Capital city  - 10 - Table 2.1(a): ANNUAL GROWTH RATE OF TOWNS IN TANZANIA 1948-57 1957-67 1948-67 Dar es Salaam 7.1 7.8 7.5 Mwanza 6.5 5.8 6.1 Tabora 2.1 3.2 2.4 Tanga 7.0 4.8 5.9 Moshi 6.1 6.9 6.5 Musoma 10.4 7.9 9.1 Mbeya 9.1 6.1 7.5 Morogoro 6.6 5.7 6.1 Lindi 2.1 2.6 2.4 Mtwara/Mik. - 2.9 - Arusha 7.3 12.4 10.0 Dodoma 4.0 5.8 4.9 Iringa 5.9 8.5 7.3 Kigoma 2.8 Source: R.A. Henin: 1967 Population Census of Tanzania: A Demo- graphic Analysis. Bureau of Resource Assessment and Land Use Planning, Research Paper No. 19, October 1972, Dar es Salaam.  - 11 - the annual growth rate of major Tanzanian towns between the intercensal period 1948 and 1967. Although the rate of population growth in Tanzania during this period was about 3% per annum, only Tabora and Lindi grew at a lower rate of 2.4%. Other towns grew at least twice as rapidly. Eldoret and Nanyuki were towns in Kenya with annual growth rates less than that of the general population (Table 2.1(b)). The rapid rate of growth of urban centers in Africa can be attri- buted to: (a) the rate of natural increase, and (b) high rural-urban migra- tion. Most African countries have experienced very high population growth rates in this century, largely because of decline in mortality due to improved health services and the introduction of modern medicine. The demographic transition theory suggests that the decline in mortality would eventually be followed by a decline in fertility. In Africa, this has yet to happen. Crude birth rates have continued to be high. Many African countries have crude birth rates of between 48 and 50 per thousand (Table 2.2). Although one might expect that the urban fertility rate would be lower than those of rural areas--a pattern found to exist in developed countries--there is conflicting empirical evidence from Africa about this. Several studies, for example, Caldwell's Ghana study 1/ have indicated lower urban fertility in Africa in recent years. Henin 2/ found that in Tanzania in 1967 the crude birth rate in urban areas was 32.7 per thousand 1/ See for instance: J. C. Caldwell: Fertility Differentials as Evidence of Incipient Fertility Decline in a Developing Country: The Case of Ghana. Population Studies 21: 5-7. 2/ R. A. Henin: 1967 Population Census of Tanzania. A DemoRraphic Analysis. Bureau of Resource Assessment and Land Use Planning. Research Paper, No. 19, October 1972. Dar es Salaam.  - 12 - Table 2.1(b): POPULATION AND GROWTH RATES OF MAJOR TOWNS OF KENYA Population Annual Town 1962 1969 Growth Rate Nairobi 266,794 509,286 9.7 Mombasa 179,575 247,073 4.7 Nakura 38,181 47,151 3.1 Kisumu 23,526 32,431 4.7 Thika 13,952 18,387 4.0 Eldoret 19,605 18,196 0.5 Nanyuki 10,448 11,624 1.6 Kitale 9,342 11,573 3.1 Malindi 5,818 10,757 9.2 Kericho 7,692 10,144 4.0 Nyanza 7,857 10,004 3.5 a/ The comparative 1962/1969 figures have in some instances been affected by changes in the town boundaries. Sources: 1. Kenya Population Census 1962, Volume III and 1969, Volume 2, Nairobi 1966 and 1971. 2. ILO Mission - Working Table reproduced in SH Ominde. Urbanization in Africa: A Kenya Case Study: ILO Study Course on Social and Economic Development of Southern Africa.  - 13 - Table 2.2: CRUDE BIRTH RATES SELECTED AFRICAN COUNTRIES Rate of Nat. Birth Rate Crude Death Rate Increase Country 1950-55 1970-75 1950-55 1970-75 1950-55 1970-75 Ghana 51 50 28 21 23 29 Upper Volta 50 48 33 26 17 22 Ivory Coast 46 50 27 21 19 29 Senegal 48 48 29 22 19 26 Sierra Leone 45 43 27 21 18 22 Liberia 45 50 29 20 16 30 Kenya 48 49 21 17 27 32 Tanzania 52 50 28 20 24 30 Ethiopia 51 49 34 24 17 25 Zaire 47 45 28 21 19 24 UN 1975: Selected World Demographic Indicators by Countries 1950 - 2000; ESA/P/WP.55 N.Y. UN.  - 14 - as against 46.7 per thousand in rural areas. Other studies have held the opposite to be true..1/ For Zaire, Romanuike 2/ showed a higher natality in urban than in rural areas. This can be attributed to the higher propor- tion of urban married women and to the more favorable age structure of the urban population since the tendency is for members of the younger age group to leave the rural areas. The improvement in urban health conditions may have stimulated fertility by increasing fecundity, reducing the risk of miscarriage and widowhood. Moreover, most urban residents are recent migrants from rural areas, with value systems similar to those of their rural counterparts. Not yet having adjusted to a situation where the probability of infant mortality is lower than in the rural areas, they have invariably tended to overproduce children in the expectation that some would die. This theory suggests that given time, urban-rural fertility differentials in most African countries would conform to the situation in developed countries. However, although fertility is high in both rural and urban areas of Africa, the rate of natural increase alone cannot account for the pheno- menal growth of urban areas in Africa in the late 1960s and 1970s. The ex- planation lies in the tremendous increase in rural-urban migration during this period. Movement of people from one area to another is not a new phenomenon in Africa. In the precolonial period, such movements were motivated by need 1/ See P. 0. Olusanya: "Modernization and the Level of Fertility in Western Nigeria", Proceedings of International Population Conference pp. 812-815. Also, Michel Francoi: "Gabon" and Joseph Boute" "Zaire", both in J. C. Caldwell. Population Growth and Socio-economic Change in West Africa. Columbia University Press, 1975. 2/ A. Romanuike: The Demography of the Republic of Congo in W. Brown et al ed. The Demography of Tropical Africa. Princeton University Press, 1961.  - 15 - for virgin land to replace exhausted farm lands or by the search of water holes. There were also coerced migrations to foreign areas following defeats in battles. During the colonial periods, the nature of migration changed but the motives were basically economic. The direction of the movement was to areas in the country experiencing socioeconomic change - newly established plantations, trading depots, etc. Post independence migrations in Africa have been predominantly rural - urban in character, although this type of migration may be two-staged, that is prospective migrants may migrate in the first instance to another rural area or to a small town before migrating to a larger entity. Data on the migrant share of total urban population are not readily available for most countries. Some information can be gleaned from recent censuses in which specific questions relating to migrant status of census respondents were asked. In Ivory Coast, for instance, 54.5% of the population of Abidjan were born outside the city and can be classified as migrants. 1/ In Accra, Ghana in 1960 and 1970 censuses, migrants into the city accounted for 52.3 and 44.9 percent respectively of the total population. In Freetown, Sierra Leone, in 1975 34.9 percent of the total population were migrants. 2/ In 1971 in Tanzania, two-thirds of the total population of urban Tanzania were migrants. The distribution of migrant/nonmigrant population for the seven towns included in the 1971 Migration and Employment Survey (NUMEIST) is given in Table 2.3(a). In all urban areas, the majority of the population were recent migrants. Only 28% of the migrants into urban Tanzania originated from 1/ K. C. Zachariah: Migration in Ivory Coast, Mimeo, IBRD 1978. 2/ C. S. Okoye: Migration in Sierra Leone, Mimeo, IBRD 1978.  - 16 - Table 2.3(a): MIGRANT/NON-MIGRANT POPULATION (X) TANZANIA Town Migrant Non-Migrant Dar es Salaam 68 32 Tanga 60 40 Arusha 76 24 Mwanzo 71 29 Tabora 60 40 Dodoma 62 38 Mbeya 74 26 Total 67 33 Source: NUMEIST 1971 Table 2.3(b): BIRTH PLACE (URBAN/RURAL) OF MIGRANTS (%) TANZANIA Town Rural Urban Dar es Salaam 73 27 Tanga 69 31 Arusha 71 29 Mwanzo 74 25 Tabora 65 35 Dodoma 70 30 Mbeya 78 22 Total 72 28 Source: NUMEIST 1971  - 17 - other urban areas (Table 2.3(b)). Addo 1/ found that in 1966, 50 percent of migrants into towns in the Accra Capital District (Ghana) were from other urban areas. Almost 49% came from rural areas. 2.2 Urbanization of Women Until fairly recently, migration from one area to another in Africa has been predominantly male. Given the historical evolution of towns and cities this is not surprising. Urban areas in Africa had their raisons d'etre as centers of communication and collecting points for locally produced goods or through development of minerals. The incentives to provide labor in such centers were specifically directed at the men. 2/ The norm was for the men to migrate to sell their labor while the women stayed behind to tend the family plots and care for the children. Living conditions in such centers of migration did not encourage migration of whole families. Before the Second World War, therefore, given the short term nature of migration, it was relatively uncommon for women to accompany their husbands, although a certain number of widows, deserted wives and other unattached women went with the men. The situation in Tanzania is illustrative of other areas in Africa experiencing rapid social change. In 1948, in Tanzania there were 260 males for every 100 females in the labor receiving areas of Tanzania. 3/ By 1971, 1/ N. Addo: Internal Migration Differentials and Their Effects on Socio- demographic Change, in J. C. Caldwell Ed: Population Growth and Socio- economic ChanRe in W. Africa. Columbia University Press 1976. 2/ For a detailed analysis of the process of urbanization of African men, see Lord Hailey: An African Survey. London OUP. 1950. 3/ P. H. Gulliver: Rural Migration in a Rural Economy. E. African Studies, No. 15, 1955.  - 18 - there were 122 males for every 100 females in urban areas. That the propor- tion of female residents in urban areas of Africa has increased dramatically can also be ascertained by looking at the sex ratios in different African countries. In Table 2.4(a) we present the sex ratios of selected urban centers of Kenya for 1962 to 1969. Between the intercensal period 1962 to 1969, each town reduced dramatically the ratio of males to females. The ratio of males to 100 females in Nairobi fell from 187 to 147. In Kisumu, the ratio fell from 150 to 124. The trend, evident in Kenya, is also present in Tanzania. Table 2.4(b) contains the proportion of males to females in urban areas for three time periods, 1948, 1957 and 1967. Between 1948 and 1967, the proportion of females to males in the population of all cities increased dramatically. For instance in 1948, there were 197 males for every 100 females in Kigoma. By 1967, the ratio had fallen to 120 per 100. The change in Dar es Salaam was not as dramatic, but was still significant--from 141 males to 123 males for every 100 females. Although time-series data on female migration into urban areas in Africa are not readily available, the data from NUMEIST can be used as indicative of the process of urbanization of women in Tanzania (Table 2.5). Whereas in Table 2.5 we find that before 1950, 33% of all urban migrants were women, by 1971 the number had increased to 54 percent of all migrants. What are the reasons for this increase in the number of females among migrants? Studies 1/ of migration behavior in Africa indicate that the move from rural to urban areas is in response to differentials in the price of 1/ M. P. Todaro: Income Expectations, Rural-Urban Migration and Employ- ment in Africa. International Labor Review, Vol. 104, July-December 1971.  - 19 - Table 2.4(a): -MALES PER 100 FEMALES - KENYA Town 1962 la 1969 /b Nairobi 187 147 Mombasa 151 139 Nakuru 150 124 Kisumu 158 126 Eldoret 145 131 Thiko 151 145 Nanyuki 152 117 Kitale 133 127 Nyeri 162 147 Gilgil 158 126 Lamu 125 92 Malindi 186 127 Athi River 213 162 Isiolo 128 112 /a S. H. Ominde, Land and Population Movements in Kenya. Nairobi, London, Heineman 1968. /b) Calculated from 1969 census figures.  - 20 - Table 2.4(b): MALES PER 100 FEMALES, URBAN AREAS TANZANIA 1948 1957 1967 Arusha 110 106 104 Dodoma 165 128 115 Iringa 121 150 102 Kigoma 197 101 120 Mtwana 128 124 121 Lindi 142 124 109 Morogoro 128 127 110 Mbeya 133 121 101 Musama - 122 107 Moshi 168 167 129 Tanga 146 137 119 Tabana 111 113 101 Mwanza 139 140 120 Dar es Salaam 141 131 123 Source: R. H. Henin; 1967 Population Census of Tanzania: A Demographic Analysis. Bureau of Resource Assessment and Land Use Planning. Research Paper No. 19. October 1972.  - 21 - Table 2.5: MIGRANT POPULATION: YEAR OF ARRIVAL BY SEX (TANZANIA) Year Male (%) Female (X) 1971 46 54 1970 49 51 1969-68 55 45 1967-66 58 42 1965-61 60 40 1960-51 64 36 1950 or before 67 33 Source: R. Sabot, Economic Development and Urban Migration: A Study of Tanzania. Clarendon Press, Oxford, 1977. Table 2.6: MIGRATION POPULATION, MAIN REASON FOR COMING TO TOWN, BY SEX TANZANIAN URBAN AREAS Reason Male (%) Female (%) Total (%) To seek employment 70 9 42 To attend school 6 2 4 To live with parents or other relatives 6 73 37 To visit 10 11 10 Other 7 4 6 Total 100 100 100 Source: NUMEIST, 1971.  - 22 - labor between rural and urban areas. In this model, the individual equili- brium generates macro imbalances as long as differentials in income persist between urban and rural areas. However, the basic assumption of migrants as economically motivated and well informed does not always apply to female migrants in Africa because, until recently, very little female migration has been directed towards jobs. Women have tended to move in response to moves by husbands or parents. The data from Tanzania in Table 2.6 is illus- trative of the motive of women migrants into urban Africa. A high proportion of male migrants, 70 percent, came to seek em- ployment. Only 9 percent of the female respondents gave this as a reason. Sixty-six percent of the female migrants came initially to join their husbands as economic dependents. This point is reinforced by Table 2.7 which indi- cates that 66 percent of the total female migrants had no source of income and only 13 percent were in wage employment. Table 2.7: URBAN MIGRANT POPULATION: SOURCES OF INCOME BY SEX (TANZANIA) Male (%) Female (%) Wage Employment 73 13 Non-wage 14 7 Non-money Income 2 14 None 11 66 Total 100 100 Source: NUMEIST, 1971  - 23 - That the rate of female migration into urban areas in Africa de- rives to a large extent from male migration is indicated by Table 2.8 on marital status of migrants into urban Tanzania by year of arrival. Since 1952 there has been a rise in the proportion of male migrants who bring their wives with them. Gulliver 1/ indicated that migrants were as a matter of official policy encouraged after the war by provision of better housing and free transportation for dependents to bring their wives with them as a method of ensuring minimum preservation of African family life. The nature of mi- gration has changed since then from being short-term and temporary to a more permanent and long-term process, necessitating moves by whole families. Table 2.8: MIGRANT POPULATION: MARITAL STATUS ON ARRIVAL BY SEX AND YEAR OF ARRIVAL (percent) (Urban Tanzania) Males Females Year Unmarried Married on Unmarried Married on of Arrival Arrival Arrival 1970-71 64 36 33 67 1964-69 67 33 25 75 1953-63 60 40 26 74 1952 and before 58 42 13 87 Source: NUMEIST, 1971. As would be expected, the proportion of the immigrants who were mar- ried on arrival is considerably higher for women than for men. Furthermore, the proportion of unmarried men on arrival remained fairly stable between 1950 and 1970. The survey indicated that 47 percent of the married migrants were 1/ op. cit.  -24 - preceded or accompanied by their wives and 15 percent of the wives came within six months of the arrival of husbands, while 12 percent came later. Only 25 percent of the married men were accompanied by wives at the time of arrival. Furthermore, unmarried migrants frequently married soon after establishing themselves. Although the foregoing supports the contention that female migration rate derives essentially from the male migration rate, Table 2.9 indicates that some women may be migrating in search of better economic opportunities just as the men. Over the years, at least in Tanzania and to a large extent in other African countries, 1/ there has been an increase in the rate of female migrants who are unmarried and migrating in response to different eco- nomic opportunities in the rural and urban areas. While the percentage of unmarried male migrants at the time of arrival had remained fairly stable between 1950-1970, the proportion of unmarried female migrants had increased from 13 percent in 1950-1952 to 33 percent in 1970-1971. The rate of migra- tion of independent females had risen faster than that of married females. The migration of independent females can be explained as a response to price differentials between urban and rural areas. The widening urban-rural income gap can be cited as an important determinant of the increased rate of migra- tion of females. 2/ In 1970, the difference between non-agricultural real wage employment and agricultural self-employment in Tanzania was three times 1/ A recent study in Kenya by ILO indicated that a large number of female urban dwellers were household heads and their motive for migrating to urban areas was essentially economic. 2/ For a discussion of rural-urban income differentials, see R. Sabot, op. cit.  - 25 - what it was in 1958, while within agriculture, the difference between self employment and wage employment was approximately four times its earlier level. 1/ Given this growing income gap between urban and rural areas, it is not surprising that to many rural dwellers the prospects of wage employment should be more appealing than self employment in rural areas and that the urban centers as the source of this employment should appeal more to them than life in the rural areas, despite the uncertainties surrounding such a move. Table 2.9: MAIN REASON FOR COMING TO TOWN: BY MARITAL STATUS (FEMALE POPULATION) Urban Tanzania 1971 (Percent) Main Reason Ever Married Never Married Employee transfer .8 2.9 School 1.2 13.3 To take up previously obtained employment .2 4.9 Seek employment 2.7 13.9 Came with relatives/parents 17.4 39.8 Came because of husband 64.4 - Came to visit 9.8 21.0 Other 3.9 4.2 Source: NUMEIST, 1971. In the Survey in Tanzania, 21.7 percent of the never-married females and 3.7 percent of the ever-married females gave the chance of employment as the main reason for migrating. The significant strides made by women in education in recent years are probably also responsible for the increase over time in the number of women migrating for economic reasons. 1/ R. Sabot, op. cit.  - 26 - Table 2.10: FEMALE MIGRANTS - MAIN REASON FOR MIGRATION, BY EDUCATION To seek To attend To be with parents To be with To visit Education employment school or other relatives husband others None 4 8 69 19 Std. 1-4 5 - 6 77 11 Std. 5-8 18 7 8 53 14 Form 1-6 plus 29 11 7 45 18 Source: NUMEIST, 1971. Table 2.10 shows that the proportion of female migrants who came to seek employment rose significantly among the better educated; 4 percent for those with no education, 5 percent, 14 percent and 20 percent for those with standard 1-4, 5-8 and forms 1-6 respectively. A possible explanation for the increase among the educated female migrants seeking work is the fact that during the 1960s, there was a steady increase in the number of women in educa- tional institutions throughout Africa. For example, in all Africa excluding the Arab States, the number of female students in educational institutions doubled between 1960 and 1970, an increase of about 7% per annum. In Kenya and Tanzania there were substantial increases in numbers of both female and male students. Over the period, female primary sc0ool enrollment ratios increased from 30 percent to 40 percent and 16 percent to 28 percent respec- tively. Given the educational selectivity of the migration process in Africa, it is not surprising that this process should also affect women. To the educated, the town or urban center represents a notion of hygiene, pur- suit of elegance, earning power and higher standards of consumption. 1/ 1/ The attractiveness of urban centers and the urban bias in the provision of social and cultural amenities have been emphasized by many writers. 1. A Callaway: "Unemployment Among African School Leavers", Journal of Modern African Studies, 1:3, 1963. 2. Afolabi Ojo: "Some Cultural Factors in the Critical Density of Population in Tropical Africa", in J. C. Caldwell and C. Okonjo, ed. The Population of Tropical Africa, Longmans, 1968.  - 27 - Furthermore, the urban economy offers the possibility of employment towards which school curriculum has been directed, as well as opportunities for upward social mobility. 1/ 2.3 Summary It has been demonstrated that although urban migration in Africa has been predominantly male, in the 1960-1970 decade there was a sharp increase in the proportion of female urban migrants, which has continued. In Tanzania the proportion of migrants who are women rose from 33 percent of all those migrating before 1950 to 54 percent by 1971. The analysis indicates that while the proportion of female migrants who came as economic dependents remained high, the pull of the city as a source of employment and education contributed to the increase in the number of female migrants. The resulting change in the sex ratio completely changed the social structure of the cities. The influx of such large numbers of women into the urban eco- nomy has great implications for labor utilization. 1/ The importance of the urban area offering the opportunity for upward social mobility and its importance as a push factor in the migration process has been emphasized by: a. P. Kilby: "Industrial Relations and Wage Determination, The Failure of Anglo-Saxon Model, Journal of Developing Areas (July 1967) b. C. R. Frank: "Urban Unemployment and Economic Growth in Africa", Oxford Economic Paper, 20:2, 1968. c. Barbara Ward: "Women and Technology in Developing Countries", in Impact of Science on Society, 22:1, January-March 1970, gives examples of how modern technological innovations make it possible for women who live in apparently insufferable conditions of squalor in towns still consider themselves better off than they were in their traditional setting.  - 28 - CHAPTER 3 CHARACTERISTICS AND STRUCTURE OF THE FEMALE URBAN LABOR FORCE 3.1 Introduction In 1971, a report prepared by the UN Economic Commission for Africa concluded that the traditional role of African women in economic development was neither evident nor even acknowledged in the modern sectors of agricul- ture, industry, commerce and government. 1/ To a casual observer this appears to be the case in at least as far as the urban economy is concerned. Histori- cally, as mentioned in chapter two, male migration was short term and African women were new to the urban environment. Recently, however, there has been a tendency for very few urban migrants to return to their place of birth. The introduction of minimum wages in many African countries as in Tanzania has made it easier for urban wage earners to live together in family units. In Tanzania, it had the effect of stabilizing urban residence. 2/ The 1971 Survey revealed that 52 percent of the urban migrants had been in continuous residence for three and a half years and 33 percent for more than six years. By the 1970s, therefore, African cities had ceased to be dormitories for peas- ants supplementing their subsistence income from agriculture. As discussed in Chapter 2, by the 1970s there had been a major shift in the sex ratios of most 1/ a. Report of the Regional Conference on Education, Vocational Training and Work Opportunities for Girls and Women in African Countries, Rabat, Morocco, 22-29, May 1971. b. Women: The Neglected Human Resources for African Development, Human Resource Development Division, United Nations, E.C.A., Addis Ababa, - Ethiopia, Canadian Journal of African Studies, 6:2, 1972. 2/ Sabot, op. cit.  - 29 - urban areas in Africa. In view of the high female labor force participation rate in rural Africa (generally estimated at over 80 percent) and their numerical importance (over 50 percent of the urban working population are women) we now proceed to examine the structure of the urban labor force and the labor force performance of women in the urban economy. The description of the demographic characteristics of the urban female population provides the background to understanding women's role in the urban economy of Tanzania. What characterizes urban females can be understood in one of two ways: (i) by a comparison with males who make up the bulk of the labor force; or (ii) through a description of factors unique to females which influence their decisions to enter the labor force. This section will begin by discussing the socio-economic characteristics of women in urban Tanzania and will proceed to discuss the structure of the urban female labor force. 3.2 Socio-economic Characteristics of the Urban Females 3.2.1 Demographic Characteristics (a) Age Structure Figure 3.la contains the age pyramid for the urban population. The age profile of the female population as that of males is characterized by a considerable concentration among the peak reproductive and prime working age groups between 15 and 44 years. This is as one would expect, given the age selectivity of the migration process in Tanzania (84.5 percent of the females and 83 percent of the males are of these ages). There are some differences between the sexes. The concentration in the two youngest age categories is slightly greater for females than for males; 21.9 percent of the females, but  Figure 3 - 30 Acre PYlt4tD Li) llt4iJPO6Pt<&. 2, AGcS MAE. LL 1- WA61; I.AiZNCtt 1tL. C ' 7 w.~~~14 fketqqt o+ hjw4A " &^ .