**Appendix B**


**Table A2.** Test for exogeneity assumption of the IV.

Table A2 presents an alternative test for exogeneity of the instrument. We estimate a model that tests the independence between the shock under analysis (instrument) and the vocabulary measure, conditional on all variables that account for the availability of learning resources and child's nutritional status. The estimated coefficient is not statistically significant. CI stands for "confidence interval." R-squared indicates that 36% of the variance for the vocabulary measure is explained by the shock and the rest of independent variables.

#### **Appendix C**


Table A3 Panel A presents the confidence intervals of the estimated impacts of maternal mental health on children's vocabulary at the age of 5 years obtained from Equation (2), using Ordinary Least Squares (OLS) and Instrumental Variables (IV) as estimation approaches. Panel B also presents confidence intervals obtained from the first stage of the IV estimation.
