2024-09-23-世界银行-工作中的种族同伴效应_来自巴西工人死亡的证据（英）_47页_871kb

Racial Peer Effects at Work: Evidence from Worker Deaths in Brazil

This policy research working paper examines the impact of racial composition on worker retention in Brazil, utilizing unexpected worker deaths as exogenous shocks. The main findings indicate that a reduction in the non-white share of co-workers decreases the retention of non-white incumbents, but does not affect white incumbents. The results highlight asymmetric peer effects, driven primarily by worker-initiated quits rather than layoffs. Non-white workers show reduced retention and subsequently seek new jobs with higher racial concentrations. Homophilic preferences for same-race coworkers appear to contribute to racial segregation across workplaces.

Key Aspects:

• Data & Method: Administrative employer-employee data from Brazil (RAIS), leveraging deaths as natural experiments. Peer groups defined by same occupation within establishments.
• Main Effect: Non-white peers' departure reduces non-white incumbents' retention (1.1 pp drop over three years); no significant effects on white incumbents.
• Mechanism: Effects are driven by quits, not layoffs, and incumbents move to jobs with higher non-white shares.
• Heterogeneity: Stronger effects in white-collar jobs requiring less teamwork, among newer employees, and when non-whites initially form majority groups.
• Interation: Non-white women most affected. Racial effects and gender effects are distinct, with neither systematically reducing productivity or earnings.

Conclusion:

The study underscores unbalanced peer dynamics that perpetuate racial segregation post-hiring, suggesting that diversity policies should address both hiring and retention through staff interactions.