While studies have indicated the onset of fertility transition in Nigeria as in most Sub- Saharan countries, no systematic attempt has been made to identify the factors responsible for this trend. Existing explanation tends to draw from the demographic transition theory without exploring the value of other key variables. One of these variables is the degree of preference implementation. This study explores the role of fertility preference implementation on the onset of fertility transition in Nigeria. The study estimated the extent to which couples have been able to implement their fertility preferences with a view to better understand the factors that are responsible for the fertility changes in the country. This study was based on the socio-economic and microeconomic frameworks of fertility namely: the Bongaarts et al (1984) version and its Stover’s reformulation (1998) of the proximate determinants of fertility and the Bongaarts (1993) supply-demand framework for the analysis of the determinants of fertility, respectively. The proximate determinant of fertility frameworks holds that all demographic, socioeconomic, cultural, institutional, psychological, health and environmental factors (background variables) operate through the proximate or intermediate variables to affect fertility. The Bongaarts (1993) supply-demand framework posits that fertility (F) as measured by total fertility rate is an outcome of the interaction of supply of births (natural fertility), demand for births (wanted fertility) and degree of fertility preference implementation (an index, which measures the extent to which people have been able to implement their fertility preferences). The degree of preference implementation is in turn dependent on cost of fertility regulation and that of unwanted childbearing. The husbands’ family planning attitude and desired number of children were incorporated into the latter framework to recognize and bring out the crucial roles of the males in eventual fertility outcomes. The 1990, 1999 and 2003 Nigeria Demographic and Health Survey (NDHS) data sets for men, women and couples were used. The methodology of the three NDHS is not too different from each other. They are nationally representative cross-sectional survey of women between the ages of 15 and 49 in 1990, 1999 and 2003 and men between 15 and 64 in 1999 and 15-59 in the 2003 survey. Twenty-four focus group discussions were also conducted among the sexes across the country to gain better insight and understanding into the issues examined. Focus group research is based on facilitating an organized discussion with a group of individuals selected because they were believed to be representative of some class. The discussion is used to bring out insights and understandings in ways, which cannot be captured by questionnaire. The focus group discussions (FGD) are also national in scope. Participants were drawn from the Northern, South Eastern and South Western regions of the country. The study population consisted of 8,781; 8,199 and 7,620 women aged 15-49 interviewed during the 1990, 1999 and 2003 NDHS, respectively. In addition, 2,584 men aged 15-64 and 2,346 men aged 15-59 interviewed during the 1999 and 2003 NDHS, respectively were used. One thousand, one hundred and sixty-eight (1,168) couples’ records derived from the 2003 NDHS and 1,280 (constructed) couples’ records for 1999 were also used. A total of eighty-nine (89) men and eighty-five (85) women participated in the twenty-four (24) focus group discussion sessions. To achieve the research objectives, a wide range of analysis was carried out in the study. In the estimation of all means and medians, the Kaplan Meier survival analysis is used. Some other estimations were done using specific formulations developed for the purpose. Thirty-six binary logistic regression models were also fitted to bring out spousal influences on each other’s attitude to family planning and desired number of children. Lastly, the focus group discussion sessions were analysed by themes to give better understanding into the issues examined. Five hypotheses were tested in the study. There was a general decline in the age specific fertility rates (ASFR) between 1990 and 2003 in all the age groups. This is reflected in the total fertility rate (TFR) for the country as a whole, which declined from 6.32 in 1990 to 5.82 in 2003. The national average masks large variations in the fertility levels between subgroups in the country. The women in the North generally have higher number of children than their Southern counterparts. This is particularly marked in the early childbearing years. Fertility levels are also higher among rural residents compared to those in the urban area. Substantial differences equally exist in the fertility levels of women by their level of education with fertility being negatively associated with level of education. A comparison of the past and current fertility also confirms that fertility has been on the decline in the country. Age at first and last births have been declining and the differentials between the two show that the number of years spent in childbearing is decreasing. Non-marital birth was also found to be increasing over time but at a level below six percent and with no identifiable educational or regional pattern. Teenage motherhood is equally declining both nationally and regionally and is relatively high among teenagers from the rural area and those with less than secondary level of education. The proportion of women that progress from one parity to another decreases as parity increases and no socially imposed optimum number of children is observed (although there is a political four-child policy in existence) among the Nigerian women. Apart from age of mother at the birth of child, which has a positive association with median length of birth interval and the surviving status of preceding child (which is understandably shorter if the preceding child is dead), length of birth interval by other characteristics shows no significant variation across sub groups. Age at first marriage remained between 16 and 17 over the years. This is lower for respondents from the rural, the North and for those with less than secondary level of education. For age at first sexual intercourse, it increased over time in the age groups. Respondents from the North and rural area however initiated sexual intercourse earlier and age at first sexual intercourse increases with level of education. Among the proximate determinant indices, the index of postpartum insusceptibility has the greatest inhibiting effect, followed by that of marriage/sexually active, contraception and then sterility. In the Bongaarts model, the indices reduced total fecundity by 12.46 births in the total sample of married women in 1990; 8.90 births in 1999 and 9.45 births in 2003 while the indices jointly reduced potential fertility by 17.69 births in the total sample of sexually active women in 1990; 16.06 births in 1999 and 16.50 births in 2003 in the Stover’s reformulation. The number of children desired marginally increased over time. This could have been affected by the high proportion of non-response especially in the 1990 survey. The desired number of children is positively related to age and number of surviving children while it is negatively related to education. Number of children desired is found to be lower among urban residents and respondents from the Southern part of the country and highest among currently married women. The focus group discussion sessions also show that people have been revising the number of children they are having downward due to ‘supposed’ economic hardship in the country and the need to give quality education, training and care to the children. However, their desire remain high. The extent, to which fertility preference is achieved, is generally high and increasing over the years in the total sample of married women with some variations in the sub groups. It is higher in the urban compared to the rural; increases with level of education, lowest in the North East and highest in the South West. Surprisingly, the extent to which concordant couples achieve their fertility preferences was lower than that of discordant couples. The fitted logistic models showed no evidence of the husbands having an upper hand in the number of children desired by the wives or on their attitude to family planning and vice versa. While the husbands play greater role in fertility decision-making in households, who has the upper hand between the husbands and the wives in actual fertility outcome was not conclusive in this study. Knowledge about family planning methods and their availability is high. The costs of fertility regulation in terms of its social, economic and in particular health components as well as obstacles to the use of fertility regulation methods were highlighted. The psychological, health, social and more frequently the economic costs of unwanted childbearing were also brought out. The responses to questions on pregnancy wantedness was also validated in this study as people generally acknowledged the circumstance of the birth of additional child(ren) as accidental. Two of the five hypotheses proposed were confirmed. One, the hypothesis that ‘the degree of fertility preference implementation is higher in the south than in the North, higher among urban residents compared to the rural residents and increases as the level of education increases’ is supported by the results of the study. Two, that ‘the degree of fertility preference implementation is increasing and playing an increasing (a more positive) role in fertility changes in Nigeria’ is also supported. The study could not confirm that ‘the indices of marriage/sexual activity, postpartum insusceptibility and contraception (in that order) have the most inhibiting effect on fertility in Nigeria.’ The other two hypotheses could not be proven conclusively either. These were that ‘degree of fertility preference implementation is higher among couples with similar desired number of children than among discordant couples’ and ‘the Husbands have more influence on their wives’ family planning attitude and desired number of children than the wives have over their husbands’. These findings have programme and policy implications. For instance, although the reduction in the number of years spent in child bearing is welcome as a result of decline in the number of older women in active childbearing, the increasing entrance of women under the age of eighteen years is worrisome. This could impact negatively on school enrolment and retention as well as on the health of the women since it is recognised that women under the age of eighteen years is one of the four groups of women with higher risk of morbidity and mortality during pregnancy and childbirth. Despite the pertinent findings of this study, a number of study limitations can be identified. This include not identifying people who have achieved their desired fertility and those who have not and conducting in-depth interview with them to gain greater insight into their fertility decision-making, desire and behaviour. This aspect requires further detailed investigation. A number of programme, policy and research recommendations are made based on the findings of this study.

La transition de la fécondité en Afrique subsaharienne a débuté tardivement et ses caractéristiques diffèrent des autres régions du monde à l’instar de l’Asie et de l’Amérique Latine, d’où le qualificatif de « African exceptionalism » qui a été avancé par Caldwell et collègues et bien d’autres auteurs. Selon les estimations des Nations Unies, l’indice synthétique de fécondité en Afrique subsaharienne est passé de 6,3 en 1990 à 4,6 enfants par femme en 2021. Sa population est estimée à 1,15 milliard en 2022, passera à 2,09 milliards en 2050 puis à 3,44 milliards en 2100. L’Afrique subsaharienne serait considérée comme la locomotive de la croissance démographique mondiale selon les projections de population des Nations Unies. Beaucoup d’études s’accordent à dire que cette forte fécondité est motivée en grande partie par une forte préférence pour les familles nombreuses. Comprendre les facteurs qui influencent les préférences de fécondité est d’une importance capitale pour anticiper la demande potentielle en matière de planification familiale, évaluer les disparités de fécondité entre les pays et les régions, et anticiper les tendances futures en matière de fécondité. Dans le même temps, une large proportion de femmes en Afrique subsaharienne participe activement au marché du travail, soit 62,2 % contre 47,4 % au niveau mondial et variant de 18,2 % au Djibouti à 83,7 % à Madagascar en 2022. L’emploi est au cœur des politiques de développement et est considéré comme un catalyseur des changements de comportements en matière de reproduction. Cette thèse examine la nature des liens entre l’emploi et la fécondité, spécifiquement le désir d’enfant additionnel, afin de mieux comprendre la dynamique de la transition de la fécondité en milieu urbain. Trois articles sont élaborés en vue d’apporter des éclaircissements sur des aspects moins étudiés de cette thématique en Afrique subsaharienne. Le premier article examine comment la relation entre l'emploi des femmes et leur désir de fécondité évolue dans le temps. L’article utilise 30 enquêtes démographiques et de santé issues du milieu urbain de 15 pays qui ont au moins deux enquêtes dont la première se situe autour de l’année 1990 et la seconde après 2010. Les résultats descriptifs suggèrent qu’une baisse du désir d’enfant additionnel est observée dans neuf pays sur 15 parmi les femmes en union allant de 0,05 % au Ghana à 0,53 % en Zambie en moyenne annuelle alors que sur la même période, le taux d’emploi des femmes a augmenté dans 12 pays sur 15, variant de 0,8 % en Tanzanie à 9,8 % au Malawi. Globalement, cette étude ne trouve pas de preuve convaincante de l’effet de l’emploi des femmes sur le désir d’enfant additionnel d’enfants en milieu urbain africain, avec peu de résultats significatifs, malgré des raisons assez fortes de s'attendre à des relations claires. Le deuxième article évalue dans quelle mesure l’effet de l’emploi de la femme sur le désir d’enfant additionnel varie selon le type d’emploi du conjoint. Le type d’emploi du conjoint, souvent non pris en compte dans les études, est susceptible de modifier la relation entre l’emploi de la femme et son désir de fécondité. Pour ce faire, nous mobilisons les récentes données des enquêtes démographiques et de santé du milieu urbain de 33 pays d’Afrique subsaharienne. Les résultats des analyses multivariées de cette étude soulignent qu’il n’existe pas un effet clair et généralisé de l’emploi des femmes sur le désir d’enfant additionnel. Cependant, certaines particularités se dégagent. Dans un premier temps, l’emploi rémunéré de la femme en dehors de l’agriculture est, d’une part, négativement associé au désir d’enfant additionnel au Burundi, en Tanzanie, au Zimbabwe, et au Tchad, et d’autre part, positivement associé au désir d’enfant additionnel en Côte d’Ivoire, au Nigéria et en Sierra Léone. Dans un second temps, un effet systématique de l’emploi salarié des femmes sur le désir additionnel d’enfant n’émerge pas pour chaque type d’emploi du partenaire. Cela suppose qu’il n’existerait pas d’effet d’interaction entre l’emploi de la femme et l’emploi de l’homme. Par ailleurs, l’effet de l’emploi de la femme sur le désir d’enfant additionnel serait plus élevé que celui de l’emploi de l’homme. Le troisième article examine quant à lui le lien entre la perception des mauvaises conditions économiques dans le contexte de résidence et le désir d’enfant additionnel dans le contexte urbain du Bénin. Nous testons l’hypothèse qu’une perception négative va avoir un effet négatif sur le désir d’enfant additionnel. Pour ce faire, nous combinons, à partir des informations géospatiales détaillées des ménages (coordonnées géographiques), les données de l’enquête démographique et de santé 2017-18 du Bénin et les données Afrobaromètre de 2016 sur les enquêtes d’opinion. Les données de l’Afrobaromètre permettent de mesurer la perception des conditions économiques. Nous limitons une fois encore les analyses au milieu urbain du Bénin. Les résultats des analyses suggèrent que les femmes qui résident dans des contextes dans lesquels les perceptions des conditions économiques sont fortement négatives sont plus susceptibles de désirer un enfant additionnel. Par contre, cette association est modérée par le niveau d’éducation des femmes. Les femmes qui ont un niveau d’éducation secondaire et plus sont susceptibles de moins désirer une prochaine naissance que leurs homologues qui n'ont pas atteint le niveau d’éducation secondaire lorsqu’elles résident dans des contextes dans lesquels l’incertitude économique est élevée. La tendance s’inverse lorsque l’incertitude économique est basse. Cette thèse s’intéresse à la dichotomie entre l’emploi des femmes et leur désir d’enfant additionnel en milieu urbain africain. Elle suscite deux questionnements que nous n’avons pas pu totalement discerner et qui pourront faire l’objet de futurs travaux. En premier lieu, la faible relation ou l’inexistence de relation entre l’emploi des femmes et le désir d’enfant additionnel exprime une réalité sociologique et économique. D’un côté, l’emploi dans ces milieux n’est pas transformateur de genre pour permettre aux femmes de décider librement, de l’autre, l’emploi demeure dans le secteur informel qui permet de le concilier plus facilement avec la fécondité. En second lieu, il se peut que les données EDS ne permettent pas de mieux cerner les nuances dans la mesure de l’emploi des femmes et des hommes. Il y manque par exemple, le nombre d’heures travaillé pour mesurer l’intensité du travail, la formalisation (contrat de travail, registre comptable, bulletin de paie) pour identifier efficacement les emplois formels et informels, le revenu de l’emploi, le lieu de travail, le travail domestique et de soins ou la pluriactivité. Nous espérons néanmoins que les résultats de cette thèse vont servir à susciter le débat sur l’emploi des femmes en Afrique. / The fertility transition in sub-Saharan Africa began later and has different characteristics compared to countries in other regions, such as Asia and Latin America. This has led to the term 'African exceptionalism' as proposed by Caldwell and colleagues and other authors. According to United Nations estimates, the total fertility rate in sub-Saharan Africa decreased from 6.3 in 1990 to 4.6 children per woman in 2021. The population of the region is projected to reach 1.15 billion in 2022, 2.09 billion in 2050, and 3.44 billion in 2100. Sub-Saharan Africa is the driving force of global population growth. Many studies agree that this high fertility rate is primarily motivated by a strong preference for large families. Understanding the factors influencing fertility preferences is important for anticipating potential demand for family planning, understanding fertility gaps between countries and regions, and anticipating future fertility trends. At the same time, a large proportion of women in sub-Saharan Africa actively participate in the labor market, with 62.2% compared to the global average of 47.4%, ranging from 18.2% in Djibouti to 83.7% in Madagascar in 2022. Employment is central to development policies and a catalyst for changes in reproductive behavior. This thesis examines the nature of the links between employment and fertility, specifically the desire for additional children, to better understand the dynamics of the fertility transition in urban areas. Three articles are developed to shed light on less studied aspects of this topic in sub-Saharan Africa. The first article examines how the relationship between women's employment and their fertility desire changes over time. It uses 30 demographic and health surveys from urban areas in 15 countries, with the first survey around 1990 and the second after 2010. Descriptive results suggest a decline in the desire for additional children among married women in nine out of 15 countries, ranging from 0.05% per year in Ghana to 0.53% in Zambia. In comparison, the employment rate for women increased in 12 out of 15 countries, varying from 0.8% in Tanzania to 9.8% in Malawi. Overall, the study finds no convincing evidence of a changing effect of women's employment on the desire for additional children in urban Africa, with few significant results despite solid reasons to expect clear relationships. The second article assesses how the effect of women's employment on the desire for additional children varies depending on the husband's type of employment. The type of employment of the husband, often overlooked in studies, can influence the relationship between women's employment and their fertility desire. For this purpose, we use recent data from demographic and health surveys from urban areas in 33 sub-Saharan African countries. The results of the multivariate analyses in this study emphasize that there is no apparent and generalized effect of women's employment on the desire for additional children. However, some specific patterns emerge. Firstly, women's paid employment outside of agriculture is negatively associated with the desire for additional children in Burundi, Tanzania, Zimbabwe, and Chad but positively associated in Côte d'Ivoire, Nigeria, and Sierra Leone. Secondly, there is no systematic effect of women's salaried employment on the desire for additional children for each type of partner's employment. This result suggests that there is no interaction effect between women's and men's employment. Moreover, the impact of women's employment on the desire for additional children is higher than that of men's employment. The third article explores the relationship between perceived poor economic conditions and the desire for more children in urban Benin in western sub-Saharan Africa. We test the hypothesis that a negative perception will negatively affect the desire for additional children. To do this, we combine detailed household geospatial information (geographic coordinates) from the 2016 Afrobarometer opinion survey data with the 2017-18 Benin Demographic and Health Survey data. The results suggest that women who live in contexts where perceptions of economic conditions are strongly negative are likely to desire a child. This association is, moreover, moderated by women's level of education. Women with secondary education or higher are less likely to desire their next pregnancy compared to those who have not completed secondary education when residing in contexts with high economic uncertainty. However, this tendency is reversed when economic uncertainty is low. This thesis highlights the dichotomy between women's employment and their desire for additional children in African urban settings. It raises two questions that we have not fully answered and which may be the subject of future research. Firstly, the weak or non-existent relationship between women's employment and the desire for additional children reflects a sociological and economic reality. On the one hand, employment in these settings is not gender-transformative enough to allow women to make decisions freely; on the other hand, employment remains mainly in the informal sector, which makes it easier to reconcile with fertility. Secondly, it is possible that DHS data do not allow for a better understanding of nuances in measuring women's and men's employment. For example, it lacks the hours worked to measure work intensity and formalization (employment contract, accounting records, payslip) to effectively identify formal and informal jobs, employment income, work location, domestic and care work, or multiple job holding. Nevertheless, we hope this thesis's results will contribute to stimulating debate on women's employment in Africa.