Endogenous Stratification Equity Gains
RCTReview
Standard average treatment effect estimates in randomized trials obscure whether active labor market policies benefit the most disadvantaged job-seekers or primarily assist candidates who would have succeeded anyway.
Picture this
Endogenous stratification uses baseline demographic traits to predict each participant's baseline earning potential, then divides the sample into predicted outcome tiers. It operates like sorting hospital patients by baseline illness severity before administering medicine to verify if the sickest patients experience the greatest relative recovery.
What the evidence says
For the lowest-predicted-earnings stratum, the job application workshop increased 4-year earnings by 50% over the control mean (p = 0.0696 for equality with the high-predicted-earnings group). This intervention reduced the earnings gap between high- and low-predicted earners from 142% down to 54%.
- Who was studied
- N = 3,052 young job-seekers aged 18–29 in Addis Ababa, Ethiopia.
- How
- Split-sample endogenous stratification predicting 4-year earnings using baseline covariates, evaluated via RCT subgroup heterogeneity analysis.
What to do
Incorporate split-sample endogenous stratification into program evaluations to measure whether labor interventions systematically compress inequality between high- and low-potential workers.
From the source
"The estimated effect size for the low-predicted-earnings group is about 50% of the control mean. This causes a large reduction in earning inequality: the earning gap between the low and the high earnings group drops from 142 percent to 54 percent..."
Anonymity_or_Distance_Job_Search_and_Labour_Market_Exclusion_in.pdf