Estimation of district-level fertility using age-structured census data and assessment of spatial-socioeconomic differentials in Botswana, 2022.

backgroundReliable subnational fertility estimates are critical for evidence-based demographic policy and planning. While Botswana's national fertility decline is well documented, less is known about district-level fertility variation and the demographic and socioeconomic mechanisms underlying these patterns. This study aims to: (i) produce district-level fertility estimates using standard direct demographic procedures applied to census data; (ii) assess the consistency of direct estimates with established indirect and model-based fertility estimation techniques; and (iii) examine how women's age structure, child mortality, socioeconomic status, and marital patterns shape spatial fertility differentials across districts.

methodsThe analysis uses data from the 2022 Botswana Population and Housing Census for women aged 15-49. Direct fertility estimation was conducted through the calculation of age-specific fertility rates (ASFRs) and total fertility rates (TFRs). These estimates were systematically compared with indirect and model-based approaches, including Rele (1967), Hauer et al. (2013), Ponnapalli and Soren (2018), and Hauer and Schmertmann (2020), to assess robustness and internal consistency. Data quality was evaluated using parity distributions and reports of recent births, with attention to potential biases arising from recall error, proxy reporting, and reference-period misclassification. Descriptive demographic analysis was used to assess the influence of age structure, child mortality, women's socioeconomic characteristics, and marital composition on district-level fertility variation.

resultsDirect and model-based fertility estimates were closely aligned, with national TFR values ranging between 2.6 and 3.0 children per woman, confirming Botswana's advanced stage of fertility transition and the reliability of census-based estimation. However, substantial subnational heterogeneity was observed. Urban districts (e.g., Gaborone, Francistown, Jwaneng) exhibited below-replacement fertility, associated with higher female educational attainment, delayed entry into childbearing, and greater labour market participation. In contrast, predominantly rural districts (e.g., Ngamiland West, Barolong, Central Mahalapye) maintained higher fertility levels (approximately 3.5-4.5), linked to elevated child mortality, lower levels of women's empowerment, and persistent traditional reproductive norms. Districts undergoing socioeconomic transition displayed intermediate fertility profiles. Differences in age structure and child survival emerged as key demographic mechanisms reinforcing spatial fertility inequalities.

conclusionBy integrating standard direct demographic techniques with indirect and model-based validation, this study demonstrates the robustness of census data for subnational fertility estimation in Botswana. Persistent spatial and socioeconomic disparities highlight the need for differentiated policy responses: rural districts require intensified investments in child health, female education, and reproductive health services to sustain fertility decline, while urban areas must plan for the social and economic implications of sustained below-replacement fertility. Addressing these divergent trajectories is essential for achieving balanced and equitable demographic development in Botswana.