IshaRaut20/Malawi-cash-transfer-replication

Replication of Baird, McIntosh & Özler (2011) — estimating marriage effects of conditional vs unconditional cash transfers in Malawi using OLS and panel methods # Malawi-cash-transfer-replication Replication of Baird, McIntosh & Özler (2011) — estimating marriage effects of conditional vs unconditional cash transfers in Malawi using OLS and panel methods ### Replicating the Marriage Effects of Conditional and Unconditional Cash Transfers **Author:** Isha Raut | Boston University — MSc in Economic Policy and Practice **Course:** DS 925: Causal Inference **Date:** December 2025 --- ## Research Question Do conditional and unconditional cash transfers reduce early marriage among adolescent girls in low-income settings? This project replicates **Table VII (Columns 1 & 2)** from Baird, McIntosh & Özler (2011) — one of the most cited RCTs in development economics — using publicly available data from the Schooling, Income and Health Risk (SIHR) study in Malawi. --- ## Background The original study ran a large randomized controlled trial assigning schoolgirls and recent dropouts in Malawi to one of three groups: - **CCT** — Conditional Cash Transfer (tied to school attendance) - **UCT** — Unconditional Cash Transfer - **Control** — No transfer The authors found that both transfer programs reduced early marriage, with the conditional program having the stronger effect. Early marriage matters because it limits schooling and shapes long-term welfare outcomes for girls in low-income settings. --- ## Data - **Source:** World Bank Microdata Library — Schooling, Income and Health Risk (SIHR) Survey, Malawi - **Waves:** 2007 baseline, 2008 Round 2, 2010 Round 3 - **Sample:** 1,505 eligible girls at baseline → 1,438 in Round 2 → 1,422 in Round 3 - **Note:** Sample is smaller than the original (~2,400 girls) due to identifier structure differences in public use files --- ## Methodology - OLS regression with sampling weights - Cluster standard errors at the enumeration area (EA) level - Controls: age indicators, baseline grade, female household head, strata fixed effects - Outcome: Binary indicator for ever married by Round 2 / …