“…Table 5 provides the OLS estimates for a modified Equation (1). That is, the specification does not include mother fixed effects, α m , but controls for state and state-trend fixed effects, and for observable child-order invariant characteristics highlighted as determinants of excess female mortality in childhood in previous literature, such as mother's years of education (Oster, 2009), wealth measured by household facilities (Echavarri and Husillos, 2016;Nandi, 2015;Oster, 2009), having a brother as the eldest sibling (Bhalotra et al, 2018;Rosenblum, 2013), number of siblings (Echavarri and Husillos, 2016;Nandi, 2015;Pande, 2003), and whether the family belongs to a main religion (Bhalotra et al, 2010). Table 5 shows the estimated coefficient of the interaction term between female and SRB for the sample of all children born between 2001 and 2011 (row Full sample), the sub-sample of firstborns (row First child), the sub-sample of second children (row Second child) and the sub-sample of children born in the third or higher order (row Third or higher order).…”