← volver

SEMINARIOS ACADEMICO CEA-MIPP: «Testing for discrimination with a continuous outcome test » Nicolás Grau (FEN)

Los invitamos al ciclo de  SEMINARIOS ACADEMICO CEA-MIPP:

Cuándo: Miércoles 02 de septiembre – 13:00 PM Santiago.

Dónde: Sala 303 (ex biblioteca) – Beauchef 851, piso 3 | Departamento de Ingeniería Industrial.

Title: «Testing for discrimination with a continuous outcome test «

Speaker: Nicolás Grau (FEN)

Abstract:

The outcome test is a robust way to detect discrimination when selection is based on a predicted outcome (Becker, 1957): it compares the realized outcomes of the two groups’ marginal admits. This paper develops two notions of discrimination for continuous outcomes — risk-adjusted and risk-neutral — depending on whether the decision maker selects on a realized-performance floor or a projected-performance threshold. I show that, under the risk-adjusted notion, the outcome test can be approximated by comparing the realized outcomes of the two groups’ worst-performing admitted individuals, whereas under the risk-neutral notion, it can be approximated by correcting the two groups’ worst-performing gap by a function of the difference in their outcome noise standard deviations. Both tests are applied to first-grade primary school admissions in Chile (2014 cohort), testing for discrimination against applicants who live in a municipality other than the school’s, and to a Chilean labor-market setting, testing for discrimination against individuals with a pretrial detention history. I find strong and robust evidence of discrimination in the school-admissions setting, and weaker, notion-dependent evidence in the labor-market setting, illustrating both the practical relevance and the complementarity of the two notions.