Endogenous Stratification and Equity Gains

In Simple Words

Imagine a footrace where runner speed is obscured by heavy backpacks given to certain runners based on their background. Removing the opacity allows judges to see that those wearing the heaviest backpacks are actually fast runners, disproportionately lifting the slowest baseline performers to the front of the pack and closing the finish-line gap.

Research question

Job-seekers with disadvantaged backgrounds—such as lower education, lack of permanent work experience, or lower predicted earnings—face disproportionate labor market exclusion due to negative employer stereotypes.

Participants

Subgroup analysis of **N = 3,052 youth sample** in Addis Ababa, Ethiopia.

Study Design

Machine-learning based split-sample **endogenous stratification** (Abadie et al., 2017) categorizing participants by predicted endline earnings using baseline covariates.

Results

The application workshop generated a **467.1 ETB monthly wage increase** for the low-predicted-earnings group (50% of control mean, p < 0.01) compared to a -99.0 ETB non-significant effect for the high-predicted-earnings group (p = 0.0696 for equality), **reducing the high-to-low earnings gap from 142% to 54%**.

Takeaway

Stratify target populations by predicted baseline earnings to deliver skill-signaling tools specifically to low-predicted-earning youth.

Read alongside the research

Plain-English glossary

Original explanations, informed by the references below. Examples are illustrations, not data from the studies above. The research text stays unchanged.

Baseline

The starting measurements in a study. Researchers use these to compare what participants were like at the start with what happens later.

ExampleBefore a training programme starts, researchers measure workers’ earnings. That first measurement is the baseline.

GET-IT glossary · Baseline characteristics ↗Read more about baseline →

Related findings

• Endogenous Stratification Equity Gains →Labour & employment

  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.

• Statistical Discrimination Mitigation →Labour & employment

  Reducing signal variance through objective performance credentials reduces employer reliance on general background traits. When clear individual data is visible, decision-makers judge candidates on verified personal competence rather than group stereotypes.

• Ability-Dependent Peer Screening →Labour & employment

  Imagine trying to pick the best chess player out of a crowd. Only a master chess player has the expertise to accurately assess who among their acquaintances truly plays at an elite level, whereas a beginner cannot distinguish between a decent amateur and a grandmaster.

• Ability-Heterogeneous Wage Returns →Labour & employment

  Imagine a nightclub with a strict dress code based on exam scores where high scorers enter easily without extra effort. For high-scoring individuals, additional work experience adds little value because employers already hire them based on test scores.