Contents lists available at ScienceDirect Health & Place journal homepage: www.elsevier.com/locate/healthplace Micro-geographic targeting for precision public policy: Analysis of child sex ratio across 587,043 census villages in India, 2011 Rockli Kim a , Praveen Kumar Pathak b , Yun Xu c , William Joe d , Alok Kumar e , R. Venkataramanan f , S.V. Subramanian a,g,∗ a Harvard Center for Population and Development Studies, Cambridge, MA, USA b Department of Geography, Delhi School of Economics, University of Delhi, Delhi, India c SuperMap Software Co. Ltd, Beijing, China d Population Research Centre, Institute of Economic Growth, Delhi, India e National Institution for Transforming India (NITI), Government of India, New Delhi, India f University of Warwick, Coventry, England g Department of Social and Behavioral Sciences, Harvard T.H. Chan School of Public Health, Boston, MA, USA ARTICLEINFO Keywords: Child sex ratio Micro-geographic targeting Precision public policy Variation India ABSTRACT Child sex ratio (CSR) is a marker of disproportionate sex ratio at birth and discriminatory practices that lead to differential survival in early childhood by sex. We used the 2011 Census on rural India to present the first local analysis of CSR across 587,043 villages. In our multilevel analysis considering villages, tehsils, districts, and states/union territories, we found 96% of the total variation in CSR to be attributed to villages. About 39% of the villages were ‘boy’ areas (CSR≤88 girls per 100 boys) and another 12% had deficits in girls (88 < CSR≤93), while 11% fell in the normal range of CSR (93 < CSR≤98), another 10% had 98 < CSR≤103, and the re- maining 28% were ‘girl’ villages (CSR > 103). The magnitude of local variation in CSR was heterogeneous across states/union territories and districts. Our findings provide timely evidence to inform localized programmes like Beti Bachao, Beti Padhao to be implemented with greater precision. 1. Introduction India consistently ranks the highest among countries with the most disproportionate sex ratio at birth (SRB) (Jha et al., 2006) and excessive under-five female mortality (Guilmoto et al., 2018). In the absence of complete registration of births and deaths in India (Agnihotri et al., 2002), child sex ratio (CSR) – the proportion of girls and boys aged 0-6 years in a population – is an imperfect but important indicator to monitor sex differences in mortality. The natural SRB is approximated to be 95 girls per 100 boys with expected variation between 93 and 98 girls per 100 boys (Premi, 2001). A CSR that significantly deviates from this range reflects a combination of disproportionate sex selective abortions leading to imbalanced SRB and post-natal practices influen- cing sex differentials in child survival at early age (Mishra et al., 2009). Further, CSR is a population marker with enormous public health and developmental implications because it captures systematic social and cultural discriminatory practices that also affect access to health care, nutrition, and the general well-being of children (Hesketh and Xing, 2006). According to the Census, SRB per 100 boys in India was 90.5 girls in 2001 and 89.9 girls in 2011 (Rajan et al., 2015). Between 1961 and 2001, CSR per 100 boys aged 0–6 years worsened by six percentage point from 97.6 girls to 91.9 girls (Ministry of Health and Family Welfare, 2014; Premi, 2001), despite policy efforts to revert the long- standing trend of inequalities as well as economic development and improvements in female literacy in India (Guilmoto and Tove, 2015; Jayachandran, 2015; Kaur, 2007; Subramanian and Corsi, 2011). The continuation of current trend in sex ratio bias in children will pre- sumably lead to marriage squeeze and increased vulnerability for women (George and Dahiya, 1998). In addition to these nation-wide estimates, substantial geographic variation in sex ratio has been documented at the regional, state, and district levels in India (Agnihotri, 1996; Agnihotri et al., 2002; Bhaskar and Gupta, 2007; Chakraborty and Kim, 2010; Echávarri and Ezcurra, 2010; George and Dahiya, 1998; Larsen and Kaur, 2013; Murthi et al., 1995; Patel, 2002; Sekher and Hatti, 2010). For instance, CSR was https://doi.org/10.1016/j.healthplace.2019.02.005 Received 15 November 2018; Received in revised form 23 January 2019; Accepted 19 February 2019 ∗ Corresponding author. Professor of Population Health and Geography, Harvard Center for Population & Development Studies, 9 Bow Street, Cambridge, MA 02138, USA. E-mail address: svsubram@hsph.harvard.edu (S.V. Subramanian). Health and Place 57 (2019) 92–100 1353-8292/ © 2019 Published by Elsevier Ltd. T