Aggregate Shocks and Foundational Schooling: Evidence from India

Description This repository contains the complete replication package for the manuscript "Aggregate Shocks and Foundational Schooling: Evidence from India." The study examines how pandemic–related schooling disruption affected school non-attendance among children aged 6–14 years in India by exploiting the staggered timing of National Family Health Survey (NFHS-5) fieldwork relative to the nationwide school-closure announcement. The central hypothesis is that pandemic-era schooling disruptions altered children's educational participation, with heterogeneous effects across age groups and gender. The empirical analysis combines nationally representative household survey data from the National Family Health Survey (NFHS-4, 2015–16; NFHS-5, 2019–21), the Time Use Survey (TUS 2019 and 2024), and the Periodic Labour Force Survey (PLFS 2019–20 and 2023–24), together with harmonized district crosswalk files. The analysis uses variation in interview timing to estimate the effects of pandemic-era schooling disruption on school non-attendance and examines age-specific and gender-specific heterogeneity, reasons for school non-attendance, and complementary evidence from children's time allocation and labour-force participation. Robustness analyses include event-study models, falsification tests, propensity score methods, Oster sensitivity analysis, difference-in-differences estimators, alternative treatment definitions, alternative clustering schemes, and wild-cluster bootstrap inference. The repository includes all Stata source code, documentation, district harmonization files, intermediate datasets, final analytical datasets, and the original TUS and PLFS datasets required to reproduce the complete analysis. The NFHS raw datasets are not included because they are distributed under the Demographic and Health Surveys (DHS) Program data-use agreement. Researchers should obtain the NFHS-4 and NFHS-5 Household Member (PR) files directly from the DHS Program and place them in the prescribed directory structure before executing the replication package. The complete workflow is executed through a single master replication script (00_Master_Replication.do), which automatically prepares all analytical datasets, harmonizes district identifiers across surveys, estimates every empirical specification, reproduces all tables and figures reported in the manuscript and appendix, and generates replication log files. The replication package was developed and tested using Stata 17. The accompanying README provides detailed instructions on software requirements, folder organization, raw data sources, and replication procedures to facilitate full reproducibility and reuse.