Paying Students to Stay in School:Short- and Long-term Effects of a Conditional Cash Transfer in England

I examine the impact of the Education Maintenance Allowance, a conditional cash transfer in England that was available nationally from 2004 to 2011, on a range of short- and long-term outcomes. Average treatment effects are identified, assuming unconfoundedness, using Inverse Probability Weighting Regression Adjustment. Treatment effect heterogeneity is examined using Causal Forests, a new machine learning approach. I find beneficial impacts of EMA on retention, university attendance and, for the first time, insecure work, as measured by the probability of being on a “zero hours” contract. Other outcomes (educational attainment, risky behaviours, and labour market outcomes) are found not to be impacted